►
From YouTube: SIG Oil and Gas Community Virtual Sessions on 8th May
Description
This is the recording of the INDUS Oil & Gas Special Interest Group run on 8th May 2020
B
A
B
A
B
Basically,
if
you
want
to,
you
can
also
touch
upon.
A
B
A
B
B
A
A
B
A
A
B
A
A
D
A
B
About
itself
chat
will,
sir,
generally,
what
I
suggest
is
on
the
chat
just
give
the
name
company
name
in
the
chat
so
that
you
know
we
know.
B
D
Is
this
is
right
from
hpcl
mumbai
architecture.
D
A
You
know
yeah,
but
let's
wait
for
some.
You
know
participants.
We
have
participation
only
from
hpcl,
hml
and
sap.
Only
you
know.
A
A
C
There
are
also
cases
are
growing,
but
not
very
fast.
Now,
okay,
it's.
E
B
I
actually
happened
to
attend
a
virtual
conference
and
I
was
amazed
at
what
some
of
these
software
allows
you
to
do.
They
had
multiple
rooms,
multiple
floors,
they
had
tables,
and
you
know
you
could
decide
which
table
you
want
to
sit
in.
F
B
A
B
A
D
B
Sandeep
is
it
okay,
if
I
share
it
from
my
desktop
or
do
you
want
to
no.
A
A
A
D
F
B
Screen
just
let
me
know
if
you're
able
to
see
the
slides.
B
Okay,
so
thank
you.
Everyone
for
joining
the
first
indus
oil
and
gas
special
interest
group
meet
for
2020
for
the
next
five
or
seven
minutes.
I
think
I'll
just
quickly
introduce
what
indus
is
and
what
we
do
for
the
benefit
of
people
who
may
or
may
not
have
attended
any
of
the
previous
indus
activities.
B
B
So
here
is
a
list
of
board
of
directors
that
are
currently
active.
You
can
see
that
you
know
we
have
two
board
members
already
on
this
call
vikas
from
nyara
and
manoj's
from
hmel,
and
if
you
look
at
it,
you'll
find
people
from
across.
B
So
when
I
said,
chapters
and
sig
model
chapters
are
typically
geographically
collocated
customers
coming
together
and
talking
about
topics
which
are
common
across
all
customers.
It
doesn't
matter
which
customer
which
industry
you're
from
what
your
size
is
so,
for
example,
some
of
the
topics
that
we
cover
in
these
sessions
are
around
enterprise
support
around
licensing
around
new
product
introduction
around.
B
You
know
all
of
those
things
which
are
common
across
all
industries,
and
then
we
have
special
interest
groups,
so
these
special
interest
groups
can
fall
in
one
of
four
categories
so
right
now,
because
of
the
kobit,
I
think
we
moved
all
of
these
special
interests
to
a
monthly
or
a
virtual
format.
So
we
have
two
functional
sights
running
currently,
which
is
one
is
the
hr
special
interest
group
and
one
is
the
procurement
special
interest
group.
They
meet
every
month
a
fixed
day,
fixed
time
same
with
three
sap
solution,
specific
special
interest
group.
B
Apart
from
alm,
the
other
two
meet
on
a
monthly
basis
as
well,
then
we
have
four
industry,
specific
industry,
user
groups
that
are
running
special
interest
groups,
retail
automotive
utilities
and
oil
and
gas,
retail
and
automotive
made
up
once
in
two
months.
Oil
and
gas
is
the
first
one
utilities,
the
first
one
is
going
to
happen
next
month
and
then
we
can
decide
now
how
often
we
want
to
lead
virtually,
and
there
are
two
which
are
in
fact
three
which
are
common
across
all
customers.
One
is
driving
digital
transformation.
B
This
also
happens
every
month
fixed
day
fix
time
defeat.
The
hacker
is
also
the
security
special
interest
group
also
happens
every
month.
So
if
any
of
these
topics
are
interesting
to
you
or
if
you
have
people
in
your
team
who
might
be
interested
in
any
of
these
topics,
please
ask
them
to
join
each
one
of
these
sigs
I'll
share
the
slides
at
the
end
of
the
presentation.
You
can
actually
use
the
links
to
directly
go
ahead
and
register
right.
B
So
how
can
indus
help
its
members
so,
first
and
foremost,
they're
all
customers
of
sap
right,
so
you
can
share
experiences,
which
is
what
we're
going
to
do
today.
You
can
learn
from
sap
around
new
products
and
all
of
that,
so
that's
also
something
that
we
will
do
today
and
of
course,
overall.
The
idea
of
doing
all
of
this
is
so
that
you
can
increase
the
roi
of
your
investments
that
you've
already
made
in
sap
that
you're
already
using
that's
the
first
one.
B
Second
one
is,
since
you
are
all
using
sap
and
sap
forms
a
significant
part
of
your
id
landscapes.
It
is
important
that
you
have
a
vehicle
to
influence
sap
both
from
a
product
perspective
also
from
a
policy
perspective
also
from
a
go-to-market,
and
you
know,
strategy
perspective.
So
that
is
how
that's
also
something
that
we
do
through
multiple
channels
and,
of
course,
last
and
not
the
least.
You
also
want
to
adopt
innovation,
so
that
you
can
continue
to.
B
You
know
improve
your
business
processes
and
technology
adoption
in
your
company.
So
we
help
you
know
by
bringing
in
new
products
new
innovations
from
sap
and
partners
together,
so
that
you
can
identify
the
areas
where
you
can
actually
use
technology
to
innovate
inside
of
your
business,
so
some
guiding
principles
based
on
which
the
user
group
is
active.
So
we
are
not
just
about
one
time
and
kind
of
move
away
or
once
in
a
while
365
days
kind
of
engagement.
B
Every
month
we
have
at
least
seven
to
eight
engagements
that
we
are
running
in
some
months.
It's
even
eleven,
which
is
almost
like
every
every
week
we
are
doing
almost
two
or
three
engagements
with
the
community.
B
The
impact
of
the
community
comes
from
the
more
number
of
people
coming
together
in
the
network
and
that
kind
of
becomes
the
more
people
join.
The
more
critical
the
network
becomes
and
more
people
sharing
becomes
even
more
powerful,
the
entire
network.
So
everyone
learns
as
long
as
people
share
so
give
before
you
get
kind
of
thing
and
then,
of
course,
unity
and
diversity.
We
have
all
kinds
of
customers.
We
have
customers
who
are
using
b1
to
using
concur
to
using
whole
host
of
essay
installations
coming
together.
B
So
so
that's
probably
what
I
had
for
a
five
minute
session
in
terms
of
in
our
introduction
in
terms
of
what
indus
is,
what
it
does,
I'm
gonna
be
around.
B
So
if
anyone
has
any
questions,
just
pop
it
into
the
chat
window
and
I'll,
try
and
answer
them
and
meanwhile,
while
we
get
started
and
I
hand
it
over
to
manoj,
if
I
can
ask
everyone
to
go
to
the
chat
window
and
just
put
in
your
name,
the
company
name
and
the
city
that
you're
dialed
in
from
so
that
you
know
who
all
has
dialed
in
and
kind
of,
you
know
also
get
to
know.
Who
else
is
on
the
call.
So
thanks
everyone
for
the
opportunity
manoj.
A
A
K
A
Pushing
in
our
respective
companies,
you
know
for
digital
transformation,
but
this
particular
pandemic
has
created
a
push.
You
know
not
only
the
I.t
department
and
in
the
respective
business
department
also
to
think
beyond.
You
know
what
they
were
doing
in
routine
and
as
sap
community
in
the
special
interest
group
of
oil
and
gas,
which
is
most
active
where
we
share
each
ideas
from
each
other.
We
all
are
doing
various
projects
for
the
transformation,
so
this
platform
provides
an
opportunity
for
us
to
share
ideas
to
learn
from
each
other's
mistakes.
A
K
A
D
J
D
D
D
B
You
can
upload
it
on
the
teams
for
teams
as
well.
Oh
okay,
I
think
it's
coming
up,
maybe
yeah.
Now
we
can
see
it
good.
Okay,.
D
You
can
see
right,
okay,
so
thanks.
I
will
just
take
you
to
the
schedule
it's
from
now
it
starts
and
without
any
break
we'll
have
it
till
at
least
two
o'clock
or
maybe
215
so
quickly.
The
welcome
note
has
been
shared
by
manoj.
We
will
have
just
a
pulse
meter
after
this
we'll
just
take
you
take
your
pulse
that
how
things
are
going
on
in
the
corvette
situation,
and
after
that
we
will
have
a
session
on
innovation
in
oil
and
gas
industry.
This
will
be
taken
by
nikhil
and
money.
D
Of
course,
money
is
a
little
bit
delayed
and
so
nikhil
will
be
taking
us
through
the
innovations.
In
fact,
he
is
also
the
innovation
lead
for
sap,
followed
by
that
we
will
hear
from
hmel
a
story
on
connected
customer
b2b
supply
chain,
a
mobility
story
on
the
hmel
they
have
done
with
what
they
have
done
on
the
b2b
area.
D
After
that
from
the
customer
we
move
into
the
assets.
Let's
hear
from
raj
singh
of
sap
uae
the
story
which
he
wants
to
share
on
connected
assets
on
mobile
asset
manager.
D
We
always
wanted
to
find
it
out
that,
in
this
situation,
when
the
prices
are
quite
dipping,
how
we
can
actually
monitor
the
financial
parameters
of
a
refinery
and
the
digital
refinery
story.
The
sas
ref
case
study
is
going
to
be
narrated
by
jawad
and
huzaifah,
who
actually
were
part
of
the
implementation.
D
The
largest
one
of
the
digital
transformation
which
is
going
on
in
the
country
by
indian
oil
is
going
to
be
narrated
by
edis
strategy
on
the
strategic
is
mr
lok
khanna,
and
the
penultimate
session
is
the
analytics
and
innovation
in
oil
and
gas.
How
sap
is
trying
to
handle
this
analytics
part
of
the
advanced
analytics,
the
architecture
and
some
of
the
use
cases,
but
from
our
platform
and
technology
team,
sandeep,
rastughi
and
varun?
D
We
will,
after
that
we
will
take
a
basically
a
little
bit
pulse
that
how
the
things
went.
What
we
want
more,
what
do
you
want?
Yes,
and
when
should
we
meet
so
we'll
have
a
basically
a
pulse
meter,
and
then
we
do
the
ending
node.
As
we
see
the
sessions
are
quite
compressed
and
we
will
try
to
have
those
q
and
a
in
the
end
of
the
each
sessions.
D
D
So
what
you
do
is
that
we
will
just
try
to
take
some
of
the
quick
pulse
check
of
what
is
happening
just
to
basically
keep
everyone
on
board
and
engaged.
So
this
is
a
tool
maintenance.
So
you
can
go
to
your
browser,
www.mentee.com
and
use
the
code
836174.
D
There
are
certain
some
questions
just
for
fun:
let's
try
to
basically
see
and
get
a
basically
a.
D
D
D
A
A
D
D
D
D
D
D
J
J
D
D
D
Me
thanks
so
optimization
of
business
processes,
designing
of
maybe
with
chatbot
or
maybe
with
rpa
or
whatever
is
that
is
something
where
the
vote
has
come
maximum
so
good.
I
think
a
quick
pulse
and
at
the
end
we
will
have
again
we'll
just
like
to
hear
from
you.
How
did
it
go?
What
things
can
be
improved
and
how
soon
we
meet?
We
will
have
four
questions
so
sarang
you
can
unshare
and
now
I
request
nikhil
to
start
basically
your
presentation.
D
D
A
C
D
So,
quick
time
checks
are
15
minutes.
You
have.
C
First
of
all,
thanks
to
everyone
who
has
participated,
especially
our
customers
and
sap
colleagues,
my
topic
is
innovations
in
oil
and
gas
industry
and
I'll
share
some
insights
about
how
sap
is
helping
oil
industry
in
innovation
and
innovative
projects.
C
So,
first
of
all,
the
methodology
that
we
use
for
innovation
is
something
like
this
and
it's
based
upon
our
experiences
with
very
with
several
oil
companies
and,
of
course,
companies
from
other
industries.
C
So
this
is
the
fsdv
model,
so
whenever
we
decide
whether
or
not
to
take
an
innovation
project,
we
first
of
all
look
at
the
technical
feasibility
and
also
the
regulatory
feasibility,
for
example.
In
some
countries
the
delivery
of
petroleum
products
are
not
allowed
so
in
in
that
kind
of
situation,
we
will
have
to
avoid
that
kind
of
innovative
project.
C
Then
we
also
look
at
scalability.
Can
this
project
be
just
successful
as
a
poc,
but
not
roll
out,
or
would
it
be
successful
and
would
it
be
scalable
at
a
larger
scale?
C
C
Is
it
going
to
be
a
viable
poc
or
project,
or
will
it
just
be
an
academic
exercise?
So
these
are
the
things
that
we
look
at
fsdv
model.
I'm
sure
you
also
have
some
kind
of
stage
gate
approach
in
your
company
or
some
model
to
prioritize
the
innovation
projects,
but
please
consider
fsdv
model
or
for
the
prioritization
and
filtering
of
projects.
C
So
we've
been
working
on
innovative
projects
across
the
entire
oil
and
gas
value
chain.
There
are
just
a
few
use
cases
mentioned
here.
Actually
there
are
many
more.
This
is
just
for
a
depiction
purpose
and
we've
we've
been
working
on
two
types
of
innovation
projects.
C
There
are
some
projects
which
are
just
internally
conceptualized
by
sap
and
we
have
developed
pocs
and
there
are
some
projects
which
have
been
done
on
behalf
of
customers
or
along
with
oil
companies.
So
there
are
both
types
of
projects
available
in
this
value
chain
and
these
set
of
pocs
that
are
mentioned
here.
So
you
can
see
there
are
projects
across
upstream
on
the
project
side
in
onshore
offshore
production
pipelines,
transportation,
downstream,
retailing,
etc.
C
So,
let's
start
with
upstream
there's
a
u.s
based
independent
company
upstream
independent
company,
with
whom
we
worked
on
this
machine
learning
based
project
to
predict
the
build
up
of
paraffin
in
oil
and
gas
wells.
So,
as
you
know,
especially
our
colleagues
who
are
from
upstream
industry,
they
know
that,
based
upon
the
crude
characteristics,
especially
for
backseat
roots,
paraffin
buildup,
is
a
very
big
problem.
It
reduces
the
flow
rate,
it
also
creates
choking,
and
sometimes
it
also
causes
accidents
or
there
could
be
a
backfill
into
the
reservoir.
C
So
what
we
did
with
this
oil
company
was
that
we,
a
joint
team,
was
formed
with
three
skill
sets
so
data
scientist,
team
engineering
team
and
the
programmers
team-
and
usually
we
have
noticed
that
the
oil
and
gas
industry
runs
these
new
technology
or
innovation
projects.
C
With
these
three
skill
sets
in
a
team,
so
this
team
extracted
about
90
attributes
using
the
hana
platform,
so
they
are
using
hana
platform
as
a
unified
database
for
it
and
ot
related
data,
and
these
90
type
of
data
types
were
like
flow
rate
or
pressure
build
up
and
the
characteristic
or
sa
testing
of
crude
oil
or
crude
oil.
That's
coming
out
so
lab
test,
sampling,
test,
etc.
C
So
there
are
various
types
of
data
that
they
used
and
then
we
built
up
these
machine
learning
based
algorithms
for
predictive
analytics
of
the
extent
of
paraffin
buildup.
So
this
was
a
very
interesting
poc
which
was
done
with
them.
It
actually
helped
them
in
predicting
the
build
up
and
hence
improving
the
production,
or
we
can
say
reducing
the
deferment
rates
in
their
production.
C
Secondly,
we
did
one
more
project
with
them,
which
was
the
route
or
maintenance
schedule
optimization
for
their
maintenance
technicians.
So,
as
you
know,
in
large
oil
fields,
there
are
a
hundred
sets
of
wells
and
there
are
well
completions
and
oil
collecting
stations,
etc.
So
there
are
several
types
of
maintenance
jobs
every
day
and
the
company
oil
company
wanted
us
to
help
them
in
setting
up
an
algorithm
or
a
solver
which
will
optimize
taking
into
account
several
parameters.
C
So,
for
example,
how
much
is
the
urgency
of
that
maintenance
job
depending
upon
the
production
volumes
profitability
from
that?
Well
because
they
are
measuring
profitability
at
the
well
level,
depending
upon
the
distance
that
the
maintenance
technician
will
need
to
cover.
How
well
is
the
operator
familiar
with
that?
Well
or
those
kinds
of
equipment?
What's
the
availability
of
spare
parts,
safety,
certifications
available
with
the
with
the
maintenance,
technician,
etc?
So
several
parameters
were
taken
to
optimize
and
come
up
with
this
optimized
schedule.
So
what
type
of
solutions
we
use?
C
We
mapped
it
alongside
the
s3
arcgis
solution
and
now,
every
day,
every
at
the
beginning
of
the
shift,
the
supervisor
comes
up
with
a
maintenance
schedule,
along
with
mac
for
each
maintenance
technician
and
earlier
they
used
to
have
a
face-to-face
morning.
Meeting
a
lot
of
time
used
to
get
wasted,
but
now
the
this
maintenance
schedule
for
the
day
is
pushed
to
the
mobile
device
of
each
of
the
maintenance
technician.
C
So
this
was
also
a
very
interesting
innovation
that
we've
worked
with
upstream
oil
company.
C
There
was
another
very
interesting
project
which
we
did
with
the
shell
and
it
was
also
published
in
media
quite
a
bit.
It's
called
well
enriched
facility
management.
The
objective
here
was
again
to
help
in
doing
big
data
management
along
with
analytics,
so
we
integrated
about
20
data
sources
with
related
to
the
well
data
and
the
equipment,
data,
geospatial
data
coming
from
sc
solutions
and
erp
data,
etc.
C
So
this
was
used
to
integrate
all
the
data
and
then
doing
analytics
and
comparing
the
performance
of
various
oil
fields
and
optimizing
the
production
and
the
resources
and
people
and
other
resources
which
are
deployed
in
the
oil
field.
C
There
are
some
more
details
available
in
the
slide
in
the
enacture,
which
we
will
share
this
slide
deck
with
you,
so
please
feel
free
to
read
them
in
detail
and
come
back
to
us.
If
you
have
any
questions
from
upstream
now,
we
move
to
mid
stream
and
there's
an
example
of
by
the
way.
I
forgot
to
mention
that
in
this
case
we
have
also
prepared
a
reference
architecture
for
them
so
that
shell
can
roll
it
out
to
to
their
various
locations.
C
For
example,
it
had
multi-layer
architecture,
including
the
data
capturing
and
then
data
integration
and
assimilation,
then
analytics
and
then
presentation
layer.
So
this
was
a
multi
layered
architecture
which
was
developed
for
that
coming
to
midstream.
Now
we
have
situations
where
pipeline
companies
have
thousands
of
kilometers
long
pipeline
and
it's
very
difficult
for
them
to
to
monitor
the
corrosion
situ
situation.
C
So,
even
though
we
have
telemetry
for
a
long
time
which
helps
us
in
identifying
pipeline
leakages,
etc,
but
corrosion
is,
is
slightly
more
difficult
to
assess
and
quantify.
So,
first
of
all,
this
oil
company
was
facing
a
situation
where
you
know
these
chemical
modifiers
that
we
inject
into
the
pipeline.
C
What
is
the
optimum
concentration?
There
was
no
scientific
way
for
them
to
calculate
they
were
doing
actually
based
on
the
empirical
evidence
and
age
old
knowledge.
They
also
had
lot
of
regulators,
walls,
equipment,
etc,
that
were
very
old
and
there
was
no
way
to
digitize
them
and
it
will
be
a
long
term
project
to
replace
them
by
iot,
enabled
sensors
etc,
and
they
were
also
getting
information
from
pigs.
C
So
it
was
important
for
them
to
assimilate
all
that
information
and
come
up
with
predictions
and
precisely
determining
which
section
of
the
pipeline
is
more
prone
to
corrosion.
Okay.
So
what
we
did
was
integrated
various
sources
of
data,
as
I
mentioned,
and
the
data
of
inspection
pipeline
inspection
data
that
was
coming
in
from
operational
sources,
a
lot
of
core
business
related
data
like
financial
and
maintenance
data
from
erp,
and
then
they
also
have
osi
soft
bi
historian
server.
C
So
the
data
coming
from
that
all
that
was
integrated
in
unified
hana
database
and
then
what
we
did
was
we
built
up
this
machine
learning
based
model
which
will
take
all
this
data
and
help
in
predicting
where
the
corrosion
will
increase
and,
of
course
it
will
take
time
to
for
this
model
to
be
trained
completely.
C
C
So
there
are
various
types
of
reports
also
which
were
created
here
along
with
this
solution
that
are
mentioned
here
and
quite
beneficial,
because
now
it's
easier
for
maintenance
and
operations.
Colleagues
to
precisely
identify
which
are
the
high
risk
areas
and
get
a
full
view
of
the
pipeline.
C
System
from
midstream
we
moved
to
downstream
or
downstream
supply
chain
management.
So
there
we
there's
an
oil
company
in
southeast
asia.
It's
a
national
oil
company
with
whom
we
worked
on
this
project
wherein
the
the
customer
demand
they
were
facing,
a
problem
where
the
customer
demand
for
fuel,
especially
at
the
gas
stations.
Petrol
pumps
keeps
fluctuating,
and
there
are
lots
of
unplanned
peak
demands
and
when
that
happens,
they
have
to
suddenly
ramp
up
their
transportation
capacity.
C
So
they
wanted
to
have
a
more
systematic
view
so
that
they
can
predict
the
the
forecast
of
fuel
demand
and
accordingly
transportation
capacity
required.
C
So
what
we
did
was
that
first
of
all,
these
dispatch
analysts
were
didn't,
have
a
centralized
database
and
any
program
which
can
help
them
in
optimizing
the
dispatch,
planning,
etc
and
also
avoid
this
out
of
stock
situation.
So
they
wanted
to
avoid
both
the
extremes,
like
any
other
oil
company.
C
So
again
we
helped
them
through
this
hana
database
and
unified.
It
integrated
it
and
took
data
from
various
sources
like
historical
demand,
which
was
coming
from
their
crm
system,
but
actually
the
crm
was
also
not
deployed
in
all
the
business
units.
So
in
some
places
it
was
data
coming
in
excel
format.
C
Then
their
sales
planning
data
sales
forecast
data,
the
depot
inventory,
data
or
terminal
inventory
data
you
can
say,
and
various
type
of
truck
masters,
which
were
anyway
outsourced.
So
it
was
coming
from
our
source,
three
pl
companies
and
then
an
machine
learning
based
algorithm
was
developed
to
model
this
daily
demand.
How
much
will
it
be
for
per
product
per
location,
etc?
This
model
is
also
under
training.
Now
it
has
undergone.
C
It's
been
trained
for
for
a
long
time
now,
and
this
algorithm
was
also
able
to
tell
them
what
will
be
the
target:
opening
stock
target,
closing
stock
forecast,
actual
targets,
openings,
closing
stock
and
opening
stock,
etc.
C
C
We
also
worked
on
a
project
with
a
russian
oil
company,
luke
oil,
which
was
leveraging
iot
for
monitoring
workers,
health
and
if
we
have
time,
I
would
like
to
show
one
video
later,
which
is
which
shows
how
much
russian
companies
are
investing
in
the
new
age
technology,
especially
from
employee
health
and
safety
point
of
view.
So
what
they
did
was
lukoil
introduced
compulsory
medical
test
before
the
shift
begins
for
all
the
workers.
Okay,
so
what
happens?
C
So
there's
a
touchscreen
thing,
so
they'll
scan
their
id
card
and
their
tests
will
begin
so
things
like
blood
pressure
and
temperature
etc
are
measured.
They
also
do
breath
test
to
see
if
there
is
any
alcohol
consumption
or
the
worker
is
intoxicated,
then
there's
a
digital
signature
at
the
end
of
it
and
as
soon
as
the
employee
completes
it,
this
data
is
automatically
sent
to
to
the
doctor
and
also
then
it
goes
into
the
hr
solution
for
their
shift.
C
Beginning
and
the
attendance
is
marked
for
the
beginning
of
the
shift
for
that
employee.
We
are
currently
also.
We
have
worked
on
one
prototype
with
them,
which
will
also
provide
like
real
time
data
analytics
of
how
many
employees
were
held
due
to
medical
reasons
today
in
the
shift,
so
that
the
shift
manager
can
also
immediately
plan
for
backfilling
or
replacing
those
workers.
C
Last
case
that
I
would
like
to
share
is
for
chesapeake
energy,
it's
an
upstream
company,
but
this
can
apply
for
downstream
companies
as
well.
So
I'm
sure
most
of
you
are
aware
of
how
the
chatbots
and
the
rpa
system
works
I'll
move
to
the
to
the
benefits.
Actually,
this
is
being
used
the
by
the
maintenance
technicians
on
a
drilling
rig.
C
So
when
the
maintenance
technician
is
either
on
the
rig
or
is
in
the
production
area
going
to
oil
collecting
station
or
pipeline
or
well
completions,
the
pers,
the
technician
goes
and
investigates
a
failure
on
a
a
motor
or
a
pump
etc,
and
immediately
he
or
she
can
access
the
chatbot
to
see
what
is
the
availability
of
spare
parts
for
a
repairing
purpose
based
upon
the
requirement
there.
C
The
chatbot
will
then
take
over
and
search
about
the
availability
of
stock
being
available
in
nearby
locations
and
then
once
if
the
stock
is
available,
the
material
issue
will
take
place
goods
issue
transaction.
If
not,
then
the
chat
bot
itself
will
create,
create
the
purchase
requisition,
and
if
it
is
an
an
urgent
requirement,
then
it
can
also
create
the
purchase
order
and
also
depending
upon
the
item,
price,
etc.
C
So
this
is
the
last
example
that
I
wanted
to
share
with
you
about
the
innovations
sandeep.
Do
I
have
three
minutes?
I
just
want
to
share
one
video
of
a
russian
company
working
on
the
employee
safety.
C
All
right
no
problems,
so
thank
you
very
much.
These
were
some
thoughts,
some
insights
about
the
innovations
happening
in
oil
and
gas
industry.
There
are
many
more,
but
due
to
positive
time,
I
can
share
only
some
examples,
but
we'll
stay
in
touch
and
we'll
have
subsequent
sessions
to
discuss
more
such
cases.
So
thanks
a
lot
once
again,
everyone.
D
Yeah
thanks
nikhil.
It
was
very
interesting.
Any
questions
which
you
would
like
to
take
share
ask
nikhil,
or
you
can
put
it
on
the
chat
bot
we
can.
Basically
he
can.
We
can
respond
at
a
latter
time.
Also.
D
So
next
speaker
is
ashutosh,
so
ashutosh
is
from
hmel
and
he
will
take
us
through
the
connected
customer.
So
over
to
you
ashutosh
to
connect
you,
you
may
please
connect
ashutosh,
manages
the
marketing
supply
chain
processes
in
hmel
and
he
is
going
to
speak
on
a
length
on
how
hmel
has
is
able
to
have
developed
the
b2b
customer
platform
and
speak
more
on
that.
J
J
Is
it
now
visible
my
screen.
J
So
good
morning,
all
I'll
be
covering
the
mobility
success
case
study
for
special
vested
energy
limited
for
b2b
supply
chain,
how
we
have
leveraged
the
benefit
of
this
mobility
platform
to
enable
our
entire
supply
chain
connected
among
each
other.
J
What
the
html
cloud
is,
let's
see
what
the
component
that
we
have
taught
when
we
have
defined
the
html
cloud
strategy
and
while
we
have
taken
sfp
cloud
platform,
so
just
brief
about
it
and
then
I'll
I'll
discuss
about
these
three
major
developers
that
we
have
done
at
mel
one
for
connecting
the
customer
for
connecting
the
dealers
that
data
portal,
one
is
from
transporters
transport
portal
and
another
is
for
ins,
that
is
indian
management
system,
that
is
super
activity
and
customers
for
transportation.
J
So,
to
start
with,
is
a
joint
venture
between
hp
scale
and
the
mr
energy
investment
private
limited.
This
is
the
largest
fbi
in
the
binary
and
second
largest
fdi,
investment
in
all
in
guest
structure
in
the
country,
investment
of
more
than
usd
five
billion-
and
this
is
the
largest
investment
in
punjab.
J
This
is
some
of
the
key
project.
Components
of
the
that
we
are
very
proud
of
one
is
that.
K
J
Start
with
single
point
moving
spf
that
can
handle
the
history
of
seven
kilometer
in
the
sea.
We
have
got
darkness
crude
oil
terminal.
You've
got
2017,
look
for
the
pipeline
from
it
is
the
complex
to
the
binary.
We
have
got
captive
power
plant
of
165
megawatt
and
we
have
got
to
keep
it
the
production
capacity
of
440
kt40
that
can
be
placed
as.
A
J
Before
expansion
and
we've
got
the
world
class,
the
housing
facility.
J
So
just
to
go
about
the
html
cloud
strategy.
So
when
we
have
decided
to
you
know
the
move
into
the
cloud,
our
strategy
was
to
provide
the
innovative
and
relevant
services
information
transaction
over
a
secure
platform
to
each
of
our
stakeholders,
the
supply
chain
to
my
customer,
to
my
employee,
to
my
alliances
with
this
partner
and
everyone.
J
So
before
once
we
decided
this
mobility
strategy,
we
have
to
select
a
mobile
development.
You
know
application
development
platform,
whichever
is
not
only
supporting
the
application
development
but
also
have
the
following
characteristics
like
it
must
be
application
agnostic.
It
should
support
sap,
non-sap
mnes
operation
technology
in
others.
It
must
support
all
standard
programming
languages,
it
must
be,
scalable
must
be
easy
to
maintain
and
the
last
and
the
foremost
the
digestive
model
must
cover
access
to
all
internet
and.
J
So
we
have
decided
a
mobility
mobile
strategy
based
on
this
of
the
other
key
components
like
it's
a
mobile
device.
First,
yes,
everybody
is
having
a
while
everybody
is
using
a
mobile,
so
organizations
are
now
leveraging
the
benefits
of
the
mobile
where
they
can.
You
know,
have
employees
with
being
more
productive
and
they
they
can
get
in
touch
with
the
customer
customers.
J
Without
the
mobile
operating
system,
we
have
focused
on
android
and
ios,
which
of
the
mobile
application
type
like
it
will
be
web
apps,
natives,
hybrid,
so
based
on
the
mobile
application
usage,
we
can
decide
upon
this,
mostly
rf.
What
we
have
developed
are
the
hybrids
there
hybrids
are
you
know
these
are
one
development
across
the
operating
system
and
then
it's
scattered
to
the
ios
or
the
android.
That
is
the
benefit,
but
it
has
the
little
bit.
J
You
know
the
limit
limitations
from
the
ui
perspective,
mobile
application
development
platform,
as
I
I've
told
you,
it
should
be
based
on
open
standards.
This
will
support
all
type
of
application
types.
It
should
have
the
quality
management
tool
for
resilient
development.
It
should
support
the
dev
life
cycle.
Management
and
last
is
the
mdm
mobile
device
management,
which
of
the
mobile
application
platform
is
supported.
What
are
the
security
functions
that
support
it,
and
then
they
should
have
the
integration
management
with
border
point
and
point
for
the
security.
J
So
this
is
the
major
you
know
the
components
while
deciding
the
mobile
setting
now
I'll
I'll.
Take
you
through
why
we
have
chosen
sap
platform,
as
in
our
enterprise
cloud
platform.
First
and
foremost,
it
integrates
application
in
the
cloud
and
on-premise
seamlessly
while
securing
allowing
connect
to
connect
people.
So
it
indicates
you
know
the
sap
and
non-sap
both
seamlessly,
and
whenever
the
you
know,
the
people
are
logging
on
to
my
database,
my
erp.
It
has
been
authenticated
on
cloud
itself,
so
that
is
the
the
major
advantage.
J
Ufc
user
experience
is
consistent
across
all
the
end
devices.
It
has
a
capability
to
connect
between
b2c.
We
have
the
access
to
business
process,
apis
that
are
developed
on
sohs
for
hana
arrival,
so
it
is
lying.
As
for
the
future
of
the
sap
development,
and
yes,
I
have
told
you
integration
with
non-sap
systems
and
allows
to
leverage
the
non-sap
data
with
this
functionality.
We
have
developed
one
of
the
most
popular
application
in
our
company,
that
is
movie
booking,
and
this
is
the
most
usually
widely
used
application
in
our
company.
J
Now,
if
I'll
talk
about
the
our
development,
the
dealer
hotel
will
talk
about.
This
is
the
completely
supply
chain
that
I
am
portraying
here.
So
the
the
sales
employee
dealers
and
the
customer
there
are
three
entities:
the
major
entities
in
the
entire
supply
chain.
If
I'll
talk
about
the
dealers
of
the
customers,
are
they
poor
to
the
supply
chain?
And
these
players
are
you
know
just
I'm
talking
about
the
polypropylene
supply
chain?
The
dealers
are
on
day-to-day
basis.
J
They
are
attracting
the
customers,
doing
many
transactions
with
them
solving
their
issues
and
and
also
creating
sales
order
on
their
behalf
for
the
supply
from
plant.
So
they
are
creating
sales
order
for
the
supply.
From
plant
or
supply
from
their
own
depose
or
supply
from
consignment
stocks
depot,
so
there
are
three
types
of
this
will
sort
of
that
the
dealers
are
performing.
J
So
this
is
the
one
of
the
transaction
up
to
here
where
they
are
creating
the
sales
order.
So
we'll
talk
about
this
is
the
one
of
the
development
deal
reporter.
So
dealer
portal
consists
of
this
functionality.
This
is
the
one
of
the
core
functionality
of
data
portal,
where
the
dealers
are
creating
sales
order.
Now
what
happened?
Next?
To
this,
this
sales
order
has
to
go
to
the
planning
for
loading.
So
what
happens
this
load?
This
sales
order
goes
to
the
load,
planning
and
sales
employee.
J
So
I
can
see
this
is
the
another
development
for
transport
photo.
This
is
the
two
major
development
to
connect
to
the
supply
chain.
What
is
happening
here?
These
portals
are
also
talking
to
each
other.
They
are
talking
to
the
sap
and
all
of
them
are
integrated
seamlessly,
whatever
the
transaction
is
happening
and
attending
each
of
the
portal
getting
updated
in
sap
in
real
time.
So,
whatever
the
information
we
are
getting
is
the
real
time
and
isn't
transparent
to
each
and
every
one
throughout
the
supply
chain.
J
So
if
I
talk
about
the
dealer
portal
so
to
enhance
the
customer
experience,
dealer
mobile
applications
develop
for
pp
dealers
and
the
sales
support,
so
we
have
developed
total
two
apps
and
one
folder
on
cloud
platform
that
has
been
listed
as
bro,
so
html
business
connect.
This
is
the
app
available
on
android
and
ios,
that
is
for
end
customer
and
channel
partner,
app
that
developed
on
android
and
ios.
That
is
what
the
dealers-
and
this
is
the
portal
with
this
financial
dot
in
I
have
shown
the
screenshot
of
the
portal.
J
Also,
this
is
the
available
portal
for
when
mainly
for
dealers,
everybody
is
the
authorized.
Person
can
knock
in
here.
So
what
all
the
functionalities
are
available
on
this
portal
talk
about
the
entire
supply
chain,
so
right
from
the
you
know,
order
management,
whether
they
can
create
or
sales
orders
from
supplies
for
supplies
from
refineries.
Yes,
on
depot,
they
can
trade.
The
real-time
status
of
the
sales
order
sales
employee
can
create
seos
for
transferring
stock
from
refinery
to
cs
depot.
They
can
track
their
sales
orders.
J
Dealers
can
process
the
sales
water
and
can
make
deliveries,
so
the
delivery
processing
transactions
are
also
provided
there.
After
this,
the
under
the
finance
section
dealers
can
create
invoices.
They
can
see
the
customer
outstanding
credit
exposure
bank
guarantee
see
all
of
these
transactions.
You
know
now
which
are
available
on
the
fingertip
of
the
dealer
earlier
they
used
to
call
the
sales
employee
and
then
may
follow
lots
of
the
outside
the
system.
Things
are
happening
to
get
the
status
now
on
a
figure
trip.
J
They
are
getting
the
all
the
status,
so
it's
a
real
time
and
yes,
partner
profile
management
customer
whenever
the
customer
is
on
board,
dealer
is
placing
a
request
for
on
holding
the
customer
and
then,
after
the
due
diligence
and
the
approval
the
customer
is
getting
registered
onto
the
port
and
some
of
the
mail
notification
functionalities
are
also
there
where
the
automatic
mailers
are
triggered,
or
dealers
can
go
to
their
particular
application
and
can
have
this
details
on
the
main.
So
this
is
the
entire
features
that
are
talked
about
for
the
dealer
portal.
J
Let's
go
to
the
transport
portal.
As
I
have
explained
you,
a
transport
photo
is
basically
available
for
the
transporters
to
place
vehicle
for
their
allocated
loads.
Now
before
this
implementation,
what
was
happening?
You
were
having
a
transport
folder,
but
that
was
running
on
platform
and
that
is
independent
from
sap.
So
what
we
were
doing,
we
were
downloading
the
data
from
sap
for
eligible
load,
uploading
there
making
planning
there
and
then
again
recycling
goes
on
again
downloading
an
essay
and
uploading
there.
So
in
that
case,
what
was
happening
there
are
the
chances
of
error.
J
Errors
were
happening,
loads
is,
being
you
know.
Physically
track
has
been
loaded,
but
again
it
is
available.
It
is
showing
as
an
available
for
placement
of
vehicle
on
portal,
so
these
two
systems
were
not
talking
after
this
development.
Now
what
has
happened?
We
have
developed
the
the
dashboard
for
polypropylene.
It
has
been
now.
You
know,
designed
for
real-time
load
allocation
24x7
without
human
independence.
We
have
run
this
system
earlier.
J
You
know
manually,
but
now
allocation
engine
has
been
set
to
run
after
each
and
half
an
hour,
and
then
it
is
doing
the
load
allocation.
As
per
the
logic
built
in
in
the
allocation
engine
real
time
without
any
human
dependence-
and
yes,
it
is
obviously
integrated
with
sap.
That
is.
J
So
I'll
talk
about
the
shift
three
minutes
more,
so
this
is
the
functionalities
that
we
have
provided
on
transport
portal
like
club,
manage
truck,
manage,
pull
up
quantity
approval,
so
this
is
the
all
the
functionalities
have
been
listed
here.
J
So
one
of
the
functionalities
that
we
have
provided
and
is
unique
is
that
the
quantity
approval
say
there
is
a
club
which
have
been
of
24
ton,
but
transporter
says:
can
I
place
a
truck
of
26
on
here,
so
that
request
is
getting
generated
on
properly
on
portal
only
and
going
to
the
sales
employee
after
the
approval,
the
sales
order
getting
changed
automatically
and
another
one
is
the
earlier
hour
approval,
something
something:
a
truck
comes
for
the
loading
early
before
it's
a
scheduled
arrival
time
then
system
is
blocking
automatically
and
then
based
on
the
approval.
J
Only
this
is
allowed
for
the
load.
This
is
the
transport
portal.
Let
me
talk
about
the
another
development
that
is
ims
so
in
this
imf.
This
is
for
the
pet
fox
supply
chain
sales.
Employees
who
are
interacting
with
the
customers
and
sales
employees
are
giving
their
them
allocations
that
there's
so
much
of
the
material.
So
this
is
great
is
available
for
you
for
this
defined
period.
J
This
is
all
happening
on
over
the
emails
and
the
own
action
scenes
now
after
development
of
this
portal.
What
what
is
happening
now,
sales
employee
is
giving
allocations
for
good
customers
for
individual
customers,
different
type
of
locations
to
the
customers,
and
now
customers
who
are
interacting
with
the
transporters,
are
placing
indent
with
the
vehicle
detail,
and
now
this
is
the
part
of
the
intent
with
vehicle
index.
That's
where
the
sensor
is
going
to
the
plms
for
loading
and
invoice.
J
So
yes,
this
is
the
system
that
implementing
implemented
allowing
marketing
sales
officer
to
allocate
materials
to
different
customers
and
traders
to
online
portals,
and
they
are
in
turn
placing
the
vehicle
request
to
leave
the
materials.
So
this
will
be
major
functionality
that
has
been
provided
as
part
of
this
portal
location
by
sales.
Employees
create
intents
by
end,
customer
intent,
tracking
and
vehicle
initiation
required.
There
are
some
of
the
benefits
like
stimulus,
information
flow
from
customer
to
the
marketing
coordination
of
the
office
staff
and
suppliers.
J
J
J
D
K
Thank
you
just
give
me
a
verbal
if
you
can
see
my
screen.
D
K
Okay,
great,
so
thank
you,
everyone
I
know.
We've
only
got
15
minutes
or
so
so
I'll
try
to
get
through
a
lot
of
slides
quite
quickly
without
speaking
to
every
word
on
the
slide.
So
just
by
way
of
introduction,
thank
you
for
the
warm
welcome.
First
of
all,
originally
from
punjab,
grew
up
in
england
and
had
the
privilege
of
working
with
customers
all
over
the
world
in
terms
of
asset
management
from
u.s,
including
putting
some
customers
even
into
antarctica.
K
So
really,
my
focus
is
on
customer
success,
so
sharing
best
practices.
That's
what
I'll
try
to
do
here
in
this
presentation
and
just
communicate
some
ideas
about
the
solution,
what
other
customers
have
experienced
and
then
obviously
looking
forward
to
taking
some
next
steps
together
with
you
as
a
follow-up?
K
So
let's
take
a
little
look
back,
I
suppose,
which
gives
a
little
bit
of
history.
I
know
that
these
in
these
sessions.
It's
always
a
lot
of
power
points,
but
I
always
try
to
include
something
of
some
humor
if
we
can-
and
this
is
a
little
bit
of
painful
humor-
I
suppose-
but
this
is
where
maintenance
is
coming
from.
K
So
the
first
generation
of
maintenance
was
really
paper
based
everything,
almost
break
fix,
running
around
white
boards,
manual
scheduling,
a
lot
of
paper
forms
and
sadly
many
of
the
customers
that
I'm
talking
to
are
still
doing
this,
regardless
of
having
sap.
K
So
you
know
just
bear
in
mind
and
take
a
look
where
things
stand
whilst
we're
talking
about
this.
So
this
was
the
kind
of
the
first
evolution
of
maintenance.
Next
came
the
introduction
of
cmms,
you
know
computerized
maintenance
management
solutions
and
one
of
those
was
sap
plant
maintenance,
albeit
very
good.
It
put
a
lot
of
data
together,
a
lot
of
best
practices.
Many
folks
customized
this
to
a
great
extent
and
got
value.
K
K
K
So
these
were
the
really
good
niche
solutions
like
meridian
and
others
that
delivered
an
exceptional
amount
of
you
know
roi,
but
they're
isolated
to
one
area,
and
you
know
this
has
been
the
evolution
now
what
sap
has
been
trying
to
work-
and
you
know
I'm
not
trying
to
sell
anything
here
just
to
try
to
give
you
a
kind
of
a
message
if
you
will
of
what
the
current
thinking
is,
is
really
to
provide
a
much
more
unified
approach
through
more
of
an
intelligent
asset
management.
K
So
this
is
linking
whether
it
be
the
suppliers,
the
manufacturers
you
as
the
operator,
your
service
teams,
through
collaboration
through
digital,
twin
and
nikhil,
mentioned
a
really
great
example
of
the
corrosion
and
additives
and
that
go
into
the
pipeline,
and
we
also
experience
this
with
the
biggest
oil
company
in
the
world.
I
think
you
can
guess
who
they
are,
but
they
projected
a
pipeline
average
lifespan
of
eight
years
where
the
mixing
was
going
on.
K
They
were
experiencing
1.5
years,
so
in
doing
a
3d
technical
view
of
the
pipeline
and
looking
at
where
the
corrosion
was
happening,
we,
the
objective,
is
really
to
increase
that
1.5
years
to
3
to
5,
maybe
more
right.
So
this
is
how
some
of
these
newer
technologies,
like
this
idea
of
the
digital
twin,
is
really
helping.
K
So
again,
this
is
where
firmly
in
this
fourth
generation,
if
you
will
of
asset
management,
not
to
be
confused
with
the
industry
4.0,
but
this
is
just
to
look
at
so
you
know
connecting
the
suppliers,
digital
twin,
predictive
maintenance
and
really
asset
health
indexing,
scoring
reliability,
centered
maintenance,
etc.
All
coming
together
in
a
much
more
connected
way
and
extending
it
down
into
the
field,
which
is
where
the
maintenance
topic
comes
in
and
we'll
focus
on
that
in
earnest
going
forward.
K
So
this
is
what
we've
seen
really
of
the
graduation
of
the
last
30
40
years
and
beyond
of
where
maintenance
has
developed.
So
not
just
what
needs
to
happen
or
what
has
happened,
but
what's
the
best
thing
that
can
happen
moving
forward
right.
So
with
all
of
that
in
mind,
this
is
where
we'll
focus
for
today's
session.
So
that's
the
maintenance
a
little
bit
more
amusement
that
I
couldn't
resist.
K
So
one
of
the
things
I
love
to
do
is
to
really
visit
our
customers
walk
with
them
talk
with
them.
Hardhat
visibility
vest
visit
the
lines
and
really
check
how
things
are
actually
happening,
because
it's
one
thing
to
take
a
look
at
flowcharts
and
diagrams
and
have
architectural
discussions,
but
when
we
actually
go
in
the
field,
we
get
to
see
exactly
how
the
work
is
undertaken.
This
is
what
what
we
see
work:
orders
of
100
pages,
two
pages
three
pages
now
things
have
moved
on.
K
Many
folks
have
tried
mobility
for
work
orders,
which
is
great
it's
a
level
of
experience,
but
inspections
are
often
done
on
paper
still
and
the
interesting
thing
is,
nobody
can
actually
prove
that
the
inspector
was
on
that
measurement
point
and
at
the
allotted
time
right
and
then,
furthermore,
we
can't
tell
if
it
was
a
zero
or
a
six
they've
written
a
one
or
a
seven
or
four
or
a
nine,
and
all
of
this
data
goes
back
into
sap
for
decision
making.
K
Right
so
we
say
we're
doing
the
inspections.
Are
we
and
then
we
say,
we've
got
the
data
upon
to
make
the
decisions,
but
until
you
get
down
to
the
source-
and
this
only
happens
when
you
get
into
the
field-
to
discover
this
kind
of
thing
what's
actually
happening,
because
the
tool
is
only
as
good
as
the
information
that's
going
into
it
right.
K
K
So
these
are
the
kind
of
issues.
Customers
are
reporting
time
recording
again,
it
was
already
mentioned
with
the
scheduling.
But
what
we
see
here
is
time
gets
recorded
anecdotally.
If
the
paper's
too
clean
it
means
it
was
that
wasn't
completed
in
the
field.
It
was
completed
afterwards,
so
that
again
confirms
that
the
information
that's
been
recorded
is
not
accurate
and
what
we've
seen
here,
I'm
sorry!
K
I
don't
have
the
completed
one,
because
that's
where
you
really
get
to
find
out
the
time
recording
is
done
rounded
up
one
hour
two
hours,
one
and
a
half
hours.
Don't
wanna
know
that
I
wanna
know
42
minutes,
39
minutes,
35
minutes,
because
this
will
help
me
with
my
craft
sequencing.
It
will
tell
me
that
in
42
minutes
the
next
actor
needs
to
be
there
right,
not
three
hours.
I
have
no
visibility
of
what
the
actual
wrench
time
is
right.
K
So
this
is
the
kind
of
background,
and
then
we
get
things
like
no
barcoding
or
handwritten
labels
right
and
if
we
barcoded
everything
or
qr-coded
or
put
certain
kinds
of
tags,
this
could
actually
give
us
greater
insights.
This
is
on
the
spare
parts
and
assemblies
but
think
about
it
on
the
measurement
points
on
the
actual
functional
locations
and
the
assets.
Of
course,
there's
no
value
in
doing
this
unless
we
have
mobile.
K
But
I'm
sorry
to
say,
but
from
the
it
perspective
our
job
is
done
because
we've
deployed
mm
or
wms,
but
when
we
actually
explore
what
you
see.
Is
that
there's
a
choke
point
right?
So
all
of
these
things
really
come
in
into
play,
and
this
is
how
mobility
can
help
planning
again,
we've
seen
all
kinds
of
systems,
hopefully
we're
not
in
this
situation,
where
the
files
represent
the
asset
and
they
moved
from
box
to
box
in
terms
of
status
and
the
customer
I
visited.
K
We
talked
to
them
about
the
mrs
and
others
have
deployed
excel
and
magnifying
glass.
But
you
know
it's
not
the
customer's
fault.
Often
it's
us.
We
don't
inform
or
learn
from
each
other
as
to
what
the
possibilities
are
so
a
little
bit
of
background
there,
and
hopefully
that
kind
of
leads
nicely
into
how
the
process
of
the
mobile
and
so
this
manual
process,
you
know
from
iw38
work
orders
get
printed,
they
get
given
to
the
filed
away,
so
the
supervisor
picks
it
up.
K
He
gives
it
to
his
team,
they
go
out
and
do
the
work.
You
know
how
this
can
get
transformed
now
into
a
mobile
process
where
the
work
order
gets
created,
go
straight
onto
the
mobile
device
of
the
supervisor
or
the
engineer
who's
going
to
do
the
work
they
go
directly
to
the
site
after
picking
up
their
parts
or
getting
the
pass
delivered
to
the
side
of
the
maintenance
scan
the
asset
perform.
The
work
make
the
notes
hit
technically
complete
in
the
field
and
before
they
even
leave
the
site.
K
That's
already
updated
in
the
back
end,
no
need
to
wait
so
data
accuracy
is
there.
Paperwork
is
eliminated,
get
much
better
visibility
of
what
the
workforce
is
up
up
to,
and
you
know
better
scheduling
and
utilization
and
you'll
see
how
this
plays
out
later.
When
we
talk
about
bp
and
others,
how
they've
actually
deployed
these?
So
this
is,
you
know
how
the
mobile
flow
looks
versus
the
manual
flow
and
the
number
of
steps
that
gets
eliminated.
K
So,
of
course,
your
processes
might
be
a
little
different.
They
might
be
a
bit
more
streamlined,
but
the
universal
truth
is
if
we
can
push
the
work
order
directly
onto
the
correct
engineer,
who
goes
and
does
the
work
and
enters
the
information
in
the
field
and
we
get
visibility.
K
That's
a
huge
uplift
in
productivity
visibility
as
well
as
control,
so
the
number
of
use
cases
is
vast
that
these
solutions
can
be
used
for
so
it's
just
one
or
two
solutions,
but
they
can
be
used
for
lots
of
different
use
cases,
work,
order,
management,
inventory,
inspections,
installations,
replacement,
signature
capture,
calibrations
grounds,
maintenance,
vehicle
maintenance,
meter
management,
rounds
and
readings,
shutdowns
and
turnarounds
gps,
esri,
etc.
All
comes
in
linear
asset
management,
it's
all
within
the
capabilities
of
these
solutions.
So
typically,
what
happens?
K
Our
customers
work
with
one
or
two
use
cases
to
start
with,
and
then
they're
able
to
expand
out
knowing
the
full
capabilities
that
other
customers
around
the
world
are
doing
and
expand
the
same
product
into
other
use
cases.
Capabilities
are
already
there
so
again
worked
with
that
same
large
oil
company
that
came
up
with
121
use
cases
for
mobile
within
the
organization.
K
All
right
so
such
is
a
scope
and
we're
more
than
happy
to
support.
You
know
the
establishment
of
a
center
of
excellence
within
each
customer,
whether
we
can
instill
some
competency
to
take
the
next
step.
So
what
are
we
talking
about
here?
So
mobile
really
will
deliver
a
full
work
package
into
the
hands
of
the
user.
It
can
go
on
any
device.
K
We
have
both
a
cloud
version
as
well
as
an
on-premise
version,
so
we
can
guide
you
towards
which
wood
solution
would
make
sense,
and
you
can
make
that
determination
quite
quickly
all
solutions
work
offline
by
default,
so
we're
not
looking
at
a
browser
or
fury
or
anything
like
that.
This
is
a
native
app
that
can
sit
on
an
intrinsically
safe
device
that
can
be
used
in
a
explosion.
Risk
area
and
those
devices
are
certified
and
you
get
everything
the
full
history
of
the
work
order
full
work
package.
K
What
needs
to
be
done
where,
when
how
long
is
it
going
to
take
which
parts
are
needed?
What
are
the
operational
steps?
Technical,
diagrams,
geolocation,
linear
asset
management,
as
well
as
good
information
from
mmwms
linking
to
mrs?
It
can
read
rfid
nfc
and
all
this
is
pre-built
capability
and
all
we're
looking
to
do
is
do
configuration
to
ensure
that
if
it's
your
kind
of
requirements
or
use
case
so
that's
mobile
asset
management
with
the
same
solution
can
also
do
inspections.
K
So
we
have
a
customer
in
qa
kuwait,
national
refining,
they've
applied
these
tags.
These
are
industrial
nfc,
it
has
a
hole
in
the
middle,
so
you
can
put
a
steel
wire
and
put
this
around
an
assembly
if
need
be,
and
all
they're
doing
is
going
to
the
mobile
with
the
mobile
device
tapping
the
nfc.
It
confirms
the
time
the
inspector
was
there.
It
confirms
that
they're
on
the
correct
measurement
point
then
they
can
enter
the
reading,
create
the
notification
in
the
field.
Take
a
picture
if
need
be
and
get
completeness
of
information
right.
K
So
this
also
covers
the
rounds
and
readings
aspect
and
inspections.
Speaking
of
inspections,
also
many
customers
extending
this
to
all
kinds
of
other
inspections,
fire
equipment,
fire
extinguishers,
fly
doors,
gantries,
even
parking
violations,
because
most
of
the
customers
that
I'm
visiting
they
want
every
car
reverse
parked
right.
So
all
of
these
safety
issues
are
coming
in
and
they're
saying:
okay,
we
complete
60
different
forms
every
day.
Can
we
mobilize
this
right?
So
unless
we
start
talking
to
the
right
stakeholders,
we
are
just
coming
up
with
good
solutions.
K
Looking
for
an
application,
then
here's
another
great
example.
So
this
was
a
water
company.
What
they
did
was
they
deployed
a
mobile
solution
with
s3.
This
is
an
actual
screenshot
of
the
mobile
solution,
with
esri
embedded
into
the
maps.
You
can
see
here
the
network
you
can
see
here
the
work
order
locations
and
then
they
can
get
deployed
now.
What
they
did
was
five
different
use.
Cases
with
mrs
the
project
started
in
september.
K
Go
live
was
december,
9th,
okay,
so
mrs
esri
integration,
linear
asset
management,
geo,
enablement,
five
different
use
cases
on
work
managers.
So
the
point
is
the
solutions
are
quite
quick
to
deploy,
so
the
roi
starts
coming
out
quite
quickly,
okay
and
then
last
but
not
least,
inventory
management
mobile
apps
that
you
can
use
in
the
yard.
K
Instead
of
having
stuff
in
the
central
stores
so
again,
great
use
cases
few
more
additions
that
we
can
do
here,
such
as
the
3d
visual
enterprise,
to
bring
to
live
technical
diagrams,
where
we
can
link
individual
parts
with
the
material
numbers,
even
anime,
videos
etc,
and
this
works
its
way
into
the
mobile,
where
we
can
actually
see
the
technical
diagrams
brought
to
life.
So
this
is
interactive
bitmap.
K
K
These
technical
diagrams
can
also
come
from
that
collaboration
tool
that
I
mentioned
earlier
on
the
asset
intelligence
network,
where
we
can
pull
this
directly
from
the
manufacturer
and
put
it
into
sap
and
then
all
the
way
down
to
the
mobile
devices.
So
all
the
solutions
coming
together
in
a
joined
up
way,
videos
are
there
as
well
so
moving
on
very
quickly.
K
I
just
wanted
to
mention,
mrs
only
for
one
second,
simply
because
it's
a
really
good
tool,
a
bit
basic
and
it's
out
of
the
box
shape,
but
when
our
customers
really
get
to
start
deploying
it,
what
they
can
also
deploy
is
the
map
in
your
assets,
the
location
of
the
engineers,
the
location
of
the
assets.
This
could
actually
be
a
plant
diagram
if
you
wish
and
then
the
work
orders
here
now
when
they're
linked
to
mobile.
K
K
Yep
sure
no
problem,
look,
let
me
just
wrap
up.
Let's
take
a
look:
okay,
two
minutes
only
and
so
two
more
slides.
I
guess
so
I'll
fly
through
this.
So
we've
got
both
on-premise
version,
so
asset
manager
is
only
in
the
cloud
ios
android
work
manager
can
be
deployed
on-premise
as
well
as
in
the
cloud,
so
there's
choices
to
be
made
in
terms
of
solutions,
other
customers
who
have
deployed
and
will
conclude
on
the
next
two
slides
few
customers
in
india,
gujarat,
gas
reliance
and
many
others
as
well.
K
So
it's
I
mean
we're
talking
hundreds
of
customers
here
around
the
world,
so
we
can
really
take
out
that
these
solutions
are
tried,
tested,
proven
kyle
actually
used
the
work
clearance
management
they
developed
it
on
top
of
mobile
as
well,
so
really
good
use
cases.
K
Last
two
slides
you
can
see
here
the
returns
on
mobile
30
productivity,
80
improvement
in
completion
time.
You
remember,
I
said
three
hours:
that's
not
what
I'm
interested
in.
I
want
the
42
minutes.
They
were
able
to
get
accuracy,
so
they
now
got
accurate
numbers
and
then,
last
but
not
least,
take
a
look
at
one
of
my
favorite
customers.
I've
worked
with
them
for
many
years.
K
Bp.
This
is
a
slide.
It's
an
older
one
from
the
it
department
who
looked
at
all
these
use
cases,
but
this
was
the
return
that
they
gave
back
half
an
hour
to
three
hours
per
day
per
field.
Engineer
gained
using
mobile,
lower
operating
costs.
So
it's
mentioned
earlier
on
that
the
oil
price
33
reduction
in
average
work
order,
cost
inventory,
holding
10
to
50
reduction.
K
So
now
these
solutions
are
hard
baked
into
their
enp
backbone.
So
whenever
they
take
over
a
concession
or
a
new
facility,
this
gets
deployed
as
a
standard
template.
So
that's
it
won't
take
up
any
more
of
your
time.
I'm
happy
to
follow
up.
Sandeep
with
the
team,
but
these
solutions
are
easy
to
deploy.
We've
got
a
lot
of
great
reference
ability
and
I
look
forward
to
working
with
you
in
the
future.
D
D
I
I
D
I
Just
start
right
away:
hello,
good
day,
everybody,
this
is
jawad
khan
and
I'm
joined
by
my
colleague,
joseph
reneed.
We're
going
to
be
presenting
to
you
a
use
case
of
digital
refinery
and
with
a
customer
success
story
all
right.
So,
let's,
let's
get
into
it
right
away
a
bit
of
a
background
in
terms
of
the
business
need
of
this
scenario.
I
We
we've
realized
that
we
need
to
build
more
resilience
into
our
supply
chain,
especially
with
events
like
what's
happening
in
the
world
right
now.
We
don't
know
when
the
next
black
swan,
or
even
that
will
happen,
that
will
majorly
shake
the
economies
and
change
the
supply
chain
structure
for
many
industries
and
commodities.
I
So
I
think
it's
more
than
important
than
ever,
that
we
build
resilience
into
our
supply
chain
and
the
whole
idea
is
that
we
create
buffers
wherever
they
are
required.
This
can
be
done
in
the
example
of,
let's
say,
building
more
storage
capacity.
I
But
how
do
we
do
that
and
how
does
this
scenario
fit
into
it?
I
think
the
idea
is
that
if
we
can
somehow
bring
efficiencies
into
into
our
value
chain
hydrocarbon
value
chain
and
somehow
find
a
way
to
convert
our
opecs
or
let
the
opecs
pay
forward
for
these
capital
investments,
that
will
be,
I
think,
a
good
way
to
think
about
it.
I
So
we
suggest,
or
what
the
scenario
brings
forward,
is
one
that
we
strive
for
tracking
the
hydrocarbon
molecule
all
the
way
from
wells
to
wheels,
both
in
terms
of
costs
and
quantities,
and
so
thereby
achieving
a
horizontal
or
a
horizontal
integration
of
the
hydrocarbon
value
chain,
with
refining
in
it,
and
also
we
achieve
a
vertical
integration
of
our
hydrocarbon
value
chain.
So
wherever
we
can
connect
to
non-sap
third-party
systems,
which
can
help
increase
the
accuracy
and
the
frequency
of
data
costing
data
all
quantity
data
in
our
hydrocarbon
value
chain.
I
One
of
the
key
value
drivers
that
we
see,
which
enables
this
kind
of
tracking
and
measurement
of
efficiency,
is
hydrocarbon,
processing
visibility.
So
what
does
hydrocarbon
processing
visibility
mean?
What
it
means
is
that
we
are
able
to
track
our
feedstock
quantity
and
costs
our
intermediate
quantity
and
cost
the
quantity
and
evaluation
of
our
finished
products
losses.
We
are
able
to
measure
our
plans
with
their
actual
results.
For
example,
as
we
know
in
the
refining,
we
do
produce
refine
revenue,
optimizing,
linear
programming
plans
and
typically
done
in
systems
like
asthmatic
pims.
I
But
do
we
are
we
able
to
regularly
measure
our
progress
against
those
plans
in
the
in
the
refineries,
or
is
it
only
that
at
the
end
of
the
month,
or
at
a
later
time,
we
realize
how
we
did
so?
That's
also
something
that
we
need
to
think
about.
How
can
we
sort
of
try
to
get
a
provisional
mass
balance
on
a
daily
basis
that
can
help
us
measure
our
progress
against
the
optimized
plans,
and
that
has
a
direct
impact
on
our
revenues.
I
Similarly,
having
visibility
on
our
costs
and
the
inventories
has
a
direct
impact
on
on
the
measurement
and
thereby
eventually
improvement
of
our
cost
direct
cost,
indirect
costs
and
opportunity
costs
also
hydrocarbon
processing
visibility
can
be
translated
into
looking
at
sap
as
a
manufacturing
analytics
platform,
especially
with
tools
like
you
know:
integration
tools,
like
sap
mii.
We
are
now
able
to
seamlessly
connect
data
from
non-sap
systems
back
to
sap
and
use
sap
as
the
one
version
of
the
truth
or
the
analytics
platform
that
can
combine
multiple
sources
of
data
and
produce
meaningful
analytics.
I
You
know,
one
of
the
great
examples
of
this
is
we
discussed
with
the
customer
was
a
quality
giveaway
analysis
which
has
a
huge
opportunity
cost
associated
with
it.
So
if
you're
producing
product,
which
is
greater
than
the
quality
spec
that
is
required,
then
we're
losing
some
cost
or
some
value
there,
we're
producing
overproducing
the
product
and
that
analysis
for
refiners
becomes
very
important.
I
What
is
the
quality
giveaway
cost
that
they're
incurring
on
a
regular
basis?
I
Analytics
as
well,
right
last
but
not
least,
this
kind
of
integration,
vertical
integration
connecting
non-sap
systems
back
to
sap
allows
us
for
business
process,
automation
in
terms
of
inventory
posting
directly
without
human
intervention,
from
non-sap
systems
back
to
sap,
but
also
in
other
areas
like,
for
example,
our
colleagues
were
talking
about
maintenance
and
we've
realized
the
use
case
at
a
particular
customer
where
we
integrated
the
asset,
vibration,
recording
system
back
to
sap,
and
we
were
able
to
automate
the
generation
of
notification
whenever
the
vibration
reading
went
beyond
a
certain
threshold.
I
E
Yeah,
thank
you
jawad
and
good
day
to
everyone.
So,
as
javad
mentioned,
we
need
to
achieve
the
vertical
integration
within
the
refinery.
Now,
typically,
what
we
usually
see,
if
you
see,
on
the
left
hand
side,
it
is
showing
a
picture
of
a
refinery
where
we
are
saying
that
it
is
a
single
unit.
E
We
consider
a
refinery
as
a
single
unit
where
all
the
costs
are
captured
and
then
we,
you
know
at
the
end
of
the
month,
once
this
crude
cost
and
this
conversion
cost
is
collected
on
the
at
the
refinery
level,
we
do
or
use
our
automatic
mathematical
allocations.
Majorly
based
on,
let's
say.
E
Of
the
products
produced
we
allocate
that
on
these
particular
products
now
now
this
this
is
very
common
and
a
very
natural
way
of
you
know.
Looking
at
your
refinery,
you
know
product
costing,
and
but
now
what
we
are
saying
we
have
to
move
from
a
single
unit
to
the
multi-unit
and
why
we
are
saying
that
is
is
for
the
reason
that
you
are
going
to
get
more.
You
know
a
detailed
visibility.
You
are
going
to
follow
the
molecule
and
attach
the
cost
to
it.
E
Yeah
and
if
you
can
see,
on
the
right
hand,
side
there
is
what
is
showing
as
a
multi-unit
modeling,
where
the
crude
is
going
in
into
the
refinery,
in
a
particular
your
distillation
and
from
the
from
the
distillation
unit.
It
goes
to,
let's
say,
a
stabilizer.
It
goes
to
another
unit
and
until
the
blending
blending
is
done
now
now
with
this,
it
is
not
the
only
quantitative
or
the
molecules
which
you
are
tracing,
but
also
you
are,
you
know,
associating
the
cost
to
it.
E
That
means
it
gives
us
an
opportunity
to
calculate
the
actual
cost
of
the
product
based
on
how
the
molecule
is
flowing
within
the
refinery.
Now
further,
this
you
know
helps
us
track
all
the
inventories
within
within
this
particular
unit
based
model
as
jawad
was
mentioning
about
the
mars
balancing.
So
it
is,
it
is
very
much
possible
to
see
or
see
the
mars
balancing
report
coming
out
at
the
unit
level,
as
well
as
at
the
refinery
level,
with
this
kind
of
an
approach.
E
If
we
take
n
yeah
and
then
definitely,
we
are
also
evaluating
your
intermediate
products,
which
are
carrying
the
cost
from
the
initial
unit
of
the
cdu.
Until
the
you
know,
blending
is
happening
now
it
gives
you
further.
E
You
know
opportunities
to
evaluate
the
cost,
whether
when
your
one
of
your
unit
is
down
and
you're
moving
your
let's
say
products
from
instead
of
instead
of
that
particular
pro
unit,
you
are
shifting
it
to
another
unit,
then
how
much
of
you
know
opportunity
cost
you
have
lost
our
opportunity
margins,
you
have
lost,
so
those
kind
of
visibility
would
be
possible
with
this
particular
approach.
I
Yeah,
thank
you
so
so
before
I
go
to
the
next
one,
we're
with
this
us
case,
we
are
obviously
able
to
measure
the
actual
cost
of
production
for
the
refinery.
I
We
are
able
to
track
as
if
I
was
saying
we're
able
to
track
the
cost
all
the
way
from
the
feed
stock,
which
is
accrued
to
the
final
products
and
the
blended
products,
and
also
downstream
petrochemicals,
if
they're
part
of
your
manufacturing
complex
and
also
we
have
a
far
more
sap
based
reporting
available
for
analyzing,
your
grm's,
your
gross
refinery
margins
as
well-
and
this
has
been
done
at
the
customer
success
story
that
we'll
just
come
to
next.
So
just
a
bit
of
a
detail.
Yeah.
E
Yeah
sure,
okay,
so
since
we
have
collected
all
the
data
wealth
based
on
our
unit
based
costing
approach,
we
then
tend
to
use
that
particular
data
with
our
refinery
analytics
now,
as
we
have
earlier
spoken,
that
we
are
doing
the
early
warnings
based
on
the
preventive
meant
for
the
preventive
maintenance
based
on
the
vibration
system
of
the
machines
that
that's,
that
is
a
possibility,
then
the
mass
balancing
again,
as
we
mentioned
earlier,
that
we
can
achieve
it
at
the
you
know
a
unit
level
that
can
become
that
can
come
out
of
directly
out
of
the
sap
system
without
wasting
the
time
and
doing
the
data
massaging
in
your
excel
sheets,
compiling
the
data
from
different
refinery
system.
E
Since
we
have
integrated
your
all
your
shop
floor
systems
to
the
sap
system,
that
means
the
data
data
wealth
is
so
huge
that
you
know
it
is
giving
you
the
capabilities
of
doing
all
these
kind
of
analysis
right
out
of
the
sap
system.
Yeah
java,
if
you
can
go
to
the
next.
H
E
E
Sample
where
we
are
measuring
the
kpis
for
for
the
refinery
where
we
are,
we
can
say
that
you
know
you
can
have
the
kpis
like
per
unit
cost
of
goods
sold.
You
can
have
the
complete
cost
of
sales.
You
know
index,
you
can
have
your
location
wise.
How
much
of
you
know
product
you
have
sold
what
you
know
what
revenue
have
you
earned
and
what
are?
What
are
the
margins
you
have
you
are?
E
You
have
been
earning
not
only
at
the
gross
refinery
margin
level,
but
also
you
can
go
and
drill
down
to
the
product
level
which
gives
you
you
know.
You
know
strategic
decision
making
to
expand
your
refinery
or
you
if
you
want
to
change
your
con.
The
configuration
of
your
refinery
that
those
kind
of
you
know
margin
analysis
gives
you
not
only
the
margin
analysis,
but
also
how
your
refinery
is
performing
in
terms
of
production
efficiencies.
E
You
can
see
that
where
you
have
to
expand
so
those
kind
of
capabilities
you
can
achieve
directly
out
of
the
scp
system,
once
you
have
this
particular
you
know
model
implemented.
I
Yeah,
I
think
from
our
main
topic.
This
was
it.
We
also
wanted
to
also
briefly
touch
upon
the
fact
that
if
many,
many
of
our
clients
have
associated
petrochemicals
divisions
and
we're
able
to
take
that
cost
visibility
downstream
to
the
petrochemical
products
as
well,
so
the
sap
system,
can
you
know
you
can
change
the
cost
further
down
the
refinery
products
to
the
downstream
petrochemical
products
as
well?
I
And
you
know
if
we
have
customer
cases,
for
example
in
the
u.s,
where
that
has
that
is
being
realized
as
well,
so
the
cost
flows
all
the
way
from
the
refinery
to
chemicals,
and
then
you
can.
You
can
do
a
complete
cost
analysis,
yeah
so
sandeep
now.
Should
we
open
the
questions
or
do
you
want
me
to
go
back
to
some
of
the
content?
Talk
about
the
success.
D
Story,
no,
we
are
running
short
of
time
and
I
know
that
you
know
such
a
short
time.
This
into
a
big
topic
cannot
be
given
justice.
Probably
we
may
have
a
more
of
a
detail
call
later
on.
Thanks
for
any
questions,
kindly
just
have
a
look
at
the
chat
because
there
will
be
definitely
some
question.
Yes,
this
is
breaking
a
myth
that
that
costing
can
it
cannot
be
done
in
sap.
D
Yes,
grm
cannot
be
now
monitored
in
sap,
yet
it
can
be
done
probably
in
days
to
come,
and
maybe
my
last
speaker
will
be
touching
on
how
grm
can
be
also
simulated
in
sap,
so
that
will
be
the
basically
the
topic
which
we
are
will
be
interested.
So
thanks
both
and
now
we
will
be
moving
to
interesting
session
on
advanced
analytics
modeling
in
oil
and
gas
industry,
and
the
situation
will
be
taken
by
eminent
speaker,
dr
raja
gopalan.
J
H
Okay
good
afternoon,
first
of
all,
thank
you
for
having
me
over
it's
a
pleasure
to
be
part
of
this
sig
on
oil
and
gas.
H
So
so,
when
manoj
asked
me
to
talk
about
advanced
analytics
and
oil
and
gas,
I
spent
quite
some
time
trying
to
figure
out
what
what
is
it
that
we
should
try
to
cover
in
the
half
over
today,
and
what
I
ended
up
deciding
to
do
was
to
kind
of
give
you
a
landscape
view
of
what
is
analytics
associate
looking
at
it
from
historical
development
over
the
last
20
25
plus
years
that
I
have
been
involved
in
this
field
and
then
talking
also
about
how
some
of
these
things
obviously
have
translated
to
products.
H
I
think
we
spoke
about
some
of
them
earlier
presentations
and
also
talk
about
some
of
the
new
things
that
we
researchers
are
very
excited
about.
So
with
that,
when
you
talk
about
presentations
like
these
one
of
the
first
questions
that
people
end
up
asking
me
is
when
you
talk
about
analytics,
what
is
the
difference
between
analytics
and
this?
You
know
this
long
list
you
see
here.
What
is
how
is
it
different
from
ai
or
data
science
or
deep
learning,
or
what
have
you
so
for
the
purpose
of
this
presentation?
H
I'm
going
to
adopt
this
very
simple
philosophy
that
it
doesn't
really
matter,
whether
you
call
it
analytics
or
ai
or
neural
networks,
or
what
have
you
as
long
as
it
solves
a
problem
of
importance
to
your
company
to
your
enterprise?
Then
you
can.
H
You
should
be
looking
at
it
quite
carefully,
okay.
So
when
I
started
this
journey
and
ai
in
the
1990s
early
1990s,
this
is
what
ai
and
analytics
really
was.
It
used
to
be
called
a
symbolic
reasoning
which
had
emphasis
on
knowledge,
representation,
goal
seeking,
etc.
It
used
to
be
called
by
a
word
called
export
systems.
H
I
think
this
phrase
has
fallen
out
of
labor
today,
but
many
of
the
things
that
we
talked
about
a
few
minutes
ago
really
fall
into
that
bucket
fuzzy
logic,
perhaps
a
little
bit
more
consumer
friendly
word.
You
see
many
washing
machines
being
sold
today
with
this
word,
fuzzy
logic
on
their
you
know
ads
and
then
a
bunch
of
other
terms.
So
all
of
these
words
obviously
have
been
supplanted
by
this
idea
that
that
is
something
called
the
eye.
H
Okay
and
this
translation
sort
of
started
taking
place
in
the
early
part
of
the
21st
century,
and
what
triggered
it
primarily
was
text.
H
That's
that
famous
story
about
how
google
realized
that
they
could
in
fact
make
money
for
the
first
time
in
around
2000
2001
by
mining,
the
queries
that
people
search
for
in
the
google
engine,
so
subsequently
text
got
supplemented
with
images
and
today,
obviously
you
go
to
any.
H
Website
you
would
see
ai
being
used
to
classify
things
all
the
way
from
cats
to
cat
scans
for
medical
application,
obviously
voice
and
many
of
us
interact
with
our
automation,
our
hand,
phones
or
computers
using
voice.
We
also
have
voice
as
the
output
from
your
computers
and
those
are
all
in
the
consumer
space,
but
obviously,
in
the
enterprise
space.
Your
enterprise
probably
has
walked
quite
a
bit
in
the
digitalization
road,
and
many
of
your
transactions
are
probably
online
and
then
obviously
your
facilities
are
probably
also
quite
digitized.
H
Okay.
So
if
you
look
at
it
from
this
lens,
it
would
look
like
the
ai
story
really
is
primarily
a
21st
century
revolution,
but
those
of
us
who
have
been
with
the
oil
and
gas
industry
for
as
long
as
I
have
been
know
that
this
data
story,
this
digitalization
story,
has
been
in
the
oil
and
gas
industry
for
much
much
longer
than
that.
So
the
picture,
the
top
left
is
how
a
typical
refinery
used
to
be
operated
back
in
the
1960s
and
what
you
see
on
the
right
with
that
dca
screen.
H
This
journey
started
in
the
1980s
in
many
places
at
the
bottom
left,
you
see
refinery
lps
again,
lps
have
become
the
normal
normal
way
to
plan
refinery
operations.
We
just
had
an
interesting
presentation
on
lps.
H
H
H
So
the
way
I
I
like
to
think
of
this
as
looking
at
a
company
from
the
point
of
view
of
this
inverted
pyramid.
So
what
this
is
showing
pictorially
is
that
if
you
take,
for
instance,
a
refinery
at
the
lowest
level,
you
have
equipment,
for
example,
a
heat
exchanger,
a
pump,
a
compressor
or
a
distillation
column.
H
These
are
probably
of
the
size
of
a
few
meters
and
they
are
operating
with
many
decisions
having
to
be
taken
in
the
order
of
seconds
at
the
next
level.
Where
we
talk
about
process
supervision,
we
are
talking
about
a
collection
of
these
units,
giving
you
probably
the
preheat
chain
in
the
refinery,
and
this
obviously
is
order
of
10
meters
or
wider,
and
the
decisions
at
this
scale
end
up
being
an
order
of
minutes.
H
Likewise,
we
can
go
up
each
of
these
layers
and
at
each
level
you
will
find
that
the
size
of
the
system
we
are
talking
goes
up
by
one
order
of
magnitude,
so
you
go
from
meter
at
the
bottom
most
level
to
order
of
100
meters
when
you're
talking
about
process,
optimization
planning
and
scheduling
and
looking
at
order
of
kilometers,
and
when
you
talk
about
supply
chain
management,
you're
talking
about
hundreds
of
kilometers
and
so
on.
H
So
it
turns
out
that
there
are
different
players
in
your
enterprise
who
are
trying
to
keep
each
level
of
this
pyramid
in
control
operating
safely
and
optimally,
making
money
for
the
enterprise.
So
these
players
make
decisions
all
within
their
own
remit
and
ai.
In
fact,
from
my
own
first
hand,
experience
has
been
deployed
in
every
one
of
these
levels,
at
least
for
the
last
15
years
or
so
okay.
H
H
So
let
us,
let
me
kind
of
give
you
a
a
quick
glance
in
each
one
of
those.
So
the
first
example
I
want
to
start
with
is
something
I'm
very
proud
of.
This
was
work
done
really
20
years
ago.
Okay,
the
challenge
that
was
posed
to
us
is
the
following,
so
this
particular
refinery
had
a
crude
distillation
tower
like
many
of
your
plants,
if
you
are
in
the
defending
industry
it,
the
the
cdu
makes
a
number
of
products,
the
one
that
this
company
wanted
to
focus
on
was
diesel
for
various
reasons.
H
The
challenge
that
they
had
was
there
was
no
way
to
measure
the
quality
of
diesel
on
the
fly.
They
had
a
lab
measurement,
which
typically
would
be
taken
once
a
day
or
at
best,
once
every
eight
hour
shift.
H
Okay,
so
trying
to
optimize
the
plant
when
you
don't
have
enough
data
was
causing
them
a
challenge,
and
so
this
is
where
we
used
air
technology
called
substances
to
help
them
come
up
with
an
estimation
of
the
diesel
quality
on
the
fly
which
the
operator
would
see
on
the
dcs
panel
on
a
continuous
basis,
it
will
run
and
generate
a
new
value,
new
estimated
value
of
the
diesel
quality
every
minute,
so
the
number
would
get
updated
frequently
and
therefore
the
operator
could
then
try
to
run
the
unit
trying
to
maximize
visa.
H
H
Looking
at
the
historical
data,
we
realized
that
there
are
huge
data
quality
issues
and
roughly
10
of
the
data
was
actually
usable,
but
even
this
10
of
the
data
was
adequate
to
build
a
fairly
good
quality
neural
network,
and
once
we
had
the
neural
network,
we
validated
it
offline
for
multiple
months
and
once
the
refinery
people
were
comfortable
with
the
quality
of
the
data
of
the
results
which
are
thrown
out
by
the
new
network,
they
could
decide
to
go
online
and
have
it
integrated
with
their
dcs
okay.
H
Now,
so
there
are
various
measures
of
how
accurate
this
is.
Basically,
it
was
giving
a
correlation
quotient
of
0.8,
which
was
very
similar
to
the
accuracy
that
they
were
getting
from
the
lab
sensors,
so
that
led
them
to
be
quite
comfortable
with
this.
The
point
I
want
to
highlight
here
in
this
slide
among
one
among
many,
is
that
this
is
technology
which
has
been
deployed
for
at
least
12
years
now.
Okay
and
obviously
the
plant
has
been
using
this
for
all
that
time,
they
probably
did
not
care
that
we
use
a
neural
network.
H
It
was
not
the
sexy
word
that
it
is
today,
but
it
basically
got
the
job
done.
Okay,
so
here
us
some
of
the
learnings
that
we
had
from
the
project,
one,
the
ai
part,
is
in
these
kind
of
projects,
typically
the
easier
aspect.
The
challenge
usually
comes
from
having
to.
H
The
data
and
the
data
quality,
even
though
you
may
be
collecting
huge
amounts
of
data
in
your
plant,
the
quality
of
the
data
often
is
not
so
great.
As
I
mentioned
earlier,
we
could
only
use
roughly
ten
percent
of
the
data
that
the
plant
had
connect
collected
over
a
two
year
window,
so
so
that's
an
example
of
something
that
we
did
a
while
ago.
Here
is
another
example.
I
will
not
go
too
deep
into
this,
because
I
don't
have
a
clearance
to
talk
about
this
technology
from
the
company
that
we
worked
with.
H
I
just
want
to
tell
you
what
this
is.
So
here
is
a
problem
that
operators
in
this
particular
plant
we're
facing
what
you
see
on
the
right
is
a
photo
of
a
unit-
it's
not
probably
very
evident
here,
but
this
is
a
unit
which
is
separating
bad
quality
water
from
organic
phase
that
you
would
see
in
a
upstream
oil
and
gas
industry,
okay.
So
the
challenge
is
this:
separation
of
oil
from
water
is
something
which
typically
cannot
be
measured.
H
H
Basically,
that
would
allow
us
to
get
number
of
pictures
every
minute
from
this
and
then
did
some
fairly
straightforward
image,
analysis
and
converted
the
picture
to
a
liquid
level.
Okay,
so
that
then,
basically
made
it
possible
for
this
process
to
be
or
this
unit
to
be.
Digitized
and
operators
could
therefore
do
lots
of
things
that
they
could
not
do
previously,
because
they
didn't
have
a
numerical
quantity,
quantitative
value
that
they
could
work
with.
H
Here's
a
third
example.
This
is
this
relates
to
alarm
management.
Many
of
you,
independent
of
whether
you
are
in
a
pipeline
business
or
a
refinery,
or
even
upstream.
You
know
that
this
ends
up
being
a
fairly
big
challenge
even
today,
so
in
this
particular
plant
again,
I
won't
tell
you
the
name
of
this
plant,
but
if
you
lived
in
southeast
asia,
you
probably
can
recognize
which
plant
this
is
so
it
had
also
a
chemical,
complex,
downstream.
H
H
What
you
see
on
this
plot
is
the
number
of
alarms
that
would
be
thrown
up
by
just
that
unit
alone
during
a
typical
eight
hour
shift.
Okay,
so
on
a
good
day,
they
were
hitting
roughly
three
thousand
alarms
on
a
bad
day.
That
number
could
be
order
of
twenty
thousand
okay.
Those
of
you
who
spent
any
time
in
the
control
room
know
that
these
numbers
are
just
among
us.
H
You
can't
really
expect
any
operator
to
make
any
sense
of
this
many
alarms
and
do
things
logically,
so
we
developed,
so
we
developed
a
intelligent
alarm
management
system
which
had
all
these
features.
I
will
not
dwell
too
much
into
it,
but
basically
what
the
idea
was
that
we
want
to
go
from
this
situation.
H
They
were
with
greater
than
one
alarm
per
minute,
which
was
basically
unacceptable
to
a
situation
towards
the
bottom
of
this
table
here,
to
about
one
alarm,
every
three
to
five
minutes
and
that's
basically,
what
we
could
do
and
what
was
required
was
basically
again
primarily
data
mining
of
the
alarm
logs.
H
So
here
you
see
this
bar
chart
showing
the
number
of
alarms
with
and
without
the
alarm
management
system
and
on
average
you
would
see
a
roughly
a
over
50
reduction.
Okay,
and
this
is
the
other
way
of
looking
at
it.
So
this
is
the
plant
operation,
that
unit
operation
before
we
deployed
the
technology
and
then
the
subsequent
six
months,
and
what
you
would
find
is
that
the
average
had
gone
down
by
about
50
percent
and
even
when
they
had
abnormal
events
like
the
shutdown
of
aromatic
unit.
H
Even
in
these
kind
of
situations,
these
models
could
hold
and
tremendously
help
the
operator
and
focus
on
the
important
aspects
which
are
relevant
to
safety
and
production
and
the
optimization
okay.
So
so,
as
I
said
that
we
were
before
we
deployed
the
solution,
we
are
at
this
one
for
50
seconds
kind
of
a
limit,
and,
with
this
simple
statistical
analysis
based
approach,
we
could
bring
it
down
to
roughly
one
for
three
minutes
which
was
deemed
to
be
acceptable
by
industry
standards.
H
So
this
technology
again
has
been
in
use
in
this
plant
and
also
many
other
sister
plants
of
this
company
for
the
last
20
years.
H
So
these
are
all
examples
of
taking
data
from
a
particular
unit
or
a
process.
The
data
could
be
in
the
form
of
measurements,
temperature
pressure
flow
kind
of
mission
rules.
It
could
be
alarm
logs.
It
could
be
something
else
and
converting
that
into
something
that
is
actionable
to
the
operators
either
on
the
field
or
in
the
control
room.
H
H
The
period
before
the
process
of
the
plant
settled
down
to
the
new
grade.
During
this
period
of
what
I
call
a
transition,
the
quality
of
the
product
being
produced
was
sub-optimal.
It
could
not
really
be
sold.
Okay
and
the
company
wanted
to
figure
out
if
there
is
a
way
to
reduce
these
periods
of
transitions.
H
Just
to
give
you
a
sense
of
how
big
a
impact
this
was
making
on
the
bottom
line,
roughly
2.4
percent
of
the
entire
time.
So,
several
days
worth
of
time
during
the
year,
the
unit
was
undergoing
transition.
H
The
way
to
look
at
this
table
is
that
you
would
see
that
on
the
first
two
columns,
you
will
see
numbers
like
one
and
two
start
mode
and
end
mode
which
basically
says
that
when
they
went
from
some
grade
one
to
a
grade
two
in
the
first
time
around
it
took
them.
The
last
column
shows
a
duration
of
one
hour,
20
minutes
and
when
they
did
it
the
second
time
it
took
them
the
same
transition
from
grade
one
to
grade
two
took
them
hour
and
twenty
five
minutes.
Okay.
H
H
On
the
first
occasion
it
took
them
5
hours
and
25
minutes.
The
second
occasion
took
them
9
hours
and
35
minutes
okay.
So
that
was
a
four
hour
giveaway
when
they
did
the
exact
same
transition,
the
second
time
around.
Okay,
so
clearly
if
they
could
just
simply
reproduce
what
they
had
done
on
the
first
occasion
exactly
the
same
way,
then
they
could
have
easily
saved
nearly
four
hours
of
transition
time,
which
directly
goes
to
the
profit
margin
of
this
unit.
H
So
basically,
we
developed
a
technology
which
would
help
them
do
that
basically
developed
a
operating
procedure,
which
you
call
the
golden
transition,
which
is
what
their
best
operators
were
doing
and
developed
addition
support
system
which
enabled
their
any
operator
to
follow
exactly
the
same
golden
transition
steps.
And
when
you
do
that,
then
the
improvement
can
be
quite
a
bit.
H
What
we
found
was
that
we
could
easily
reduce
the
time
taken
for
the
transition
for
all
the
transitions
over
the
course
of
six
months
from
roughly
50
45
percent
to
nearly
80
percent,
depending
on
which
grade
they
were
switching
from
so
a
fairly
large
reduction,
which
in
that
case
led
to
a
one
one
percent
increase
in
the
unit
productivity.
Okay,
all
of
this
purely
by
doing
data
analysis
on
the
process,
data
and
then
doing
various
ai
based
analytics
on
that
now.
H
And
again
we
have
done
some
work
in
that
aspect
as
well,
so
for
the
alarms,
for
example,
we
have,
in
addition,
in
trying
to
replace
the
typical
typical
tabular
representation,
which
is
what
you
would
see
on
a
typical
alarm
display
or
alarm
log.
We
have
developed
a
number
of
graphical
displays.
H
Okay,
like
you
see
at
the
bottom,
basically
motivate
me
to
motivate
motivated
by
the
gps
kind
of
displays
that
you
and
I
are
used
to
when
using
something
like
google
maps
for
driving
our
cars.
Okay.
Now,
when
you
do
these
kinds
of
displays,
it
also
allows
us
to
incorporate
more
advanced
technologies
as
you
develop
them,
for
example,
in
this
alarm
display
that
you
see
at
the
bottom
of
the
screen,
the
left,
half,
which
says
alarm
trend.
Historical
pane
is
the
typical
alarms
that
you
would
see
in
the
traditional
display.
H
What
we
have
done
in
this
particular
work
was
also
built.
A
prediction
engine
which
would
basically
give
a
operator
an
insight
into
what
is
going
to
happen
in
the
next
minute
or
so
with
their
unit,
and
what
you
see
in
this
example
shown
in
this
picture
is
that
there
are
number
of
variables
which
are
very
close
to
the
alarm
limit
and
over
the
next
few
seconds.
H
These
variables
are
likely
to
hit
the
alarm
limit.
Okay
now.
The
reason
this
is
very
useful
is
that
this
helps
the
operator
see
a
bigger
picture
of
things
which
are
going
to
happen
in
the
future
even
before
they
do
and
therefore
their
situational
awareness
increa
increases
tremendously.
H
We
also
done
a
number
of
others,
so
this
is
something
that
we
had
developed
about
nine
years
ago
for
a
field
operator
and
on
an
offshore
platform
where
some
equipment
in
the
oil
field
is
being
monitored,
and
once
the
an
abnormality
is
seen
or
foreseen
by
the
ai
system,
then
it
would
trigger
it
as
an
alarm
to
the
field
operator,
who
would
basically
see
it
on
their
mobile
handheld
devices,
so
very
similar
to
some
of
the
things
that
was
discussed
in
the
previous
presentation.
H
H
The
reason
they
came
to
us
was
they
realized
by
looking
at
the
numbers
that
this
is
a
marine
refinery.
Vlccs
will
bring
in
crude
and
they
will
be
offloaded
into
tanks
which
will
then
feed
to
the
cdu.
They
found
that
on
many
many
occasions
they
had
to
creep
their
their
vlcc's
ships,
waiting,
which
was
leading
to
large
demand
payments,
and
basically,
they
said,
is
there
something
we
could
do
as
researchers
to
help
them
improve
the
scheduling
so
that
they
don't
have
to
pay
this
dimmer
age?
H
When
we
looked
at
this
issue
and
did
some
modeling
of
the
way
this
particular
unit,
this
particular
plant
operator,
not
just
the
refinery,
but
also
the
way,
the
entire
supply
chain
which
feeds
this
refinery,
the
way
it
is
operated
we
would
we
could
find
that
the
problem
really
did
not
originate
with
the
shipping
or
the
crude
transfer.
It
actually
originated
somewhere
else.
Okay,
so
in
order
to
help
them
also
see
this,
what
we
did
as
a
first
step
was
to
kind
of
develop.
H
This
is
excel
based
supply
chain
dashboard,
which
would
just
grab
data
from
their
system
on
the
fly
to
keep
track
of
where
the
plant
is,
what
are
the
what
it's
producing,
but
also
all
the
crews
which
are
coming
in
and
all
the
products
that
are
delivered,
that
they
committed
to
the
inventory
levels
and
so
on.
So
this
is
a
I'll
show
you
two
other
screenshots.
Okay,
what
they
could
do
with
this
dashboard
was
also
see
when
disruptions
happen.
H
So
in
this
particular
case
you
will
see
that
on
the
raw
material
shipment
side
on
the
on
the
on
the
left
table,
you
see
there's
a
red
bar
which
basically
indicates
that
this
shipment
containing
two
specific
crude
parcels
was
getting
delayed
and
that
was
going
to
lead
to
a
stock
out
situation
with
one
specific
quality
of
crude
indicated
as
oman
crude
in
the
on
the
left
side,
and
this
could
lead
to
various
issues
for
them.
Okay,
now
this
deception
could
also
based
on
some
technology.
H
They
could
plan
various
interventions
so
for
in,
for
instance,
in
this
case
you
will
see
there
is
a
emergency
crude
procurement
that
is
being
shown
in
the
in
the
on
the
left
table,
and
basically
this
dashboard
would
help
them
figure
out
what
kind
of
interventions
they
could
do
evaluate
their
effect
on
their
plan
and
the
supply
chain
and
see
if
the
deceptions
could
be
overcome-
and
you
would
see
in
this
particular
case
that
this
emergency
procurement
of
two
crude
parcels
tackles
takes
care
of
the
shortage
of
oman
crude,
that
they
were
going
to
face
otherwise,
okay.
H
So
this
is
basically
the
front
end
used
by
the
supply
chain.
Planners
what's
happening
in
the
background,
are
complex,
dynamic,
simulation
models
which
basically
mimic
the
entire
operation
of
their
supply
chain,
and
these
are
sophisticated,
agent-based
models.
I
will
not
talk
about
the
technology
right
here,
but
with
these
models,
one
could
not
just
help
them
see
the
problem,
but
we
can
also
help
them.
We
could
also
help
them
design
better
ways
to
run
their
supply
chain.
H
I
have
two
different
examples
shown
here
so
in
this
particular
table.
We
are
looking
at
various
kpis
that
that
particular
refinery
was
interested
in
monitoring
at
the
bottom
is
the
profit
in
some
monetary
units.
You
will
notice
that,
in
the
original
fashion
of
running
the
refinery,
they
were
making
a
base
case
profit
of
roughly
47
units
we
could
based
on
our
anal
analysis
and
the
models,
identify
that
they
could
change
their
procurement
policy,
and
that
would
lead
to
some
small
changes
in
the
crude
procurement
cost.
H
H
This
leads
to
a
number
of
other
benefits
also,
and
it
leads
to
another
60
increase
in
the
profit.
So,
what's
lying
behind
this
kind
of
calculations
is
a
model
driven
by
our
understanding
of
their
supply
chain
and
analytics
to
help
us
feed
this
also
into
their
dashboard.
H
H
Various
companies
industries,
from
my
first
hand,
interactions
with
them,
but
also
with
based
on
various
discussions
we
had
around
the
world
with
various
players.
These
are
things
that
they
have
made.
They
have
had
custom
made
to
their
requirements
depending
on
what
is
the
problem
that
specific
plant
or
unit
sees
as
being
important?
H
H
So
these
sub
optimal
operator
actions
can
sometimes
lead
to
fairly
big
impacts.
Again,
we
are
seeing
this
firsthand,
so
this
kind
of
motivated
us
some
years
ago
to
start
looking
at
the
quality
of
thinking
of
these
decision
makers
who
are
sitting
in
front
of
the
control
panel
or
the
field
operators.
Again.
This
was
something
which
was
discussed
in
a
previous
presentation
here.
We
are
not
necessarily
talking
about
substance,
abuse
or
alcohol
consumption.
H
We
have
seen
plans
where
operators
have
had
to
put
in
a
lot
of
overtime,
leading
to
fatigue,
not
visible
to
their
supervisor
or
manager,
and
sometimes
operators
coming
in
to
the
floor,
even
though
the
training
is
perhaps
not
up
to
the
mark,
okay,
using
new
developments
in
eye
tracking
eeg
and
so
on.
You
see
pictures
of
the
kind
of
devices
we're
talking
about.
These
are
non-invasive,
the
operator
can
wear
it.
H
We
are
able
to
generate
collect
data
about
the
mental
state
of
the
operator,
and
once
we
are
able
to
collect
data
again,
I
will
not
go
into
details
of
the
analytics.
One
can
start
answering
the
questions
that
I
have
post
here.
One
can
ask:
is
the
operator
really
focus
on
the
task,
or
is
he
distracted?
H
Is
the
operator
consistently
getting
overloaded
and
losing
the
situation,
awareness
which
is
important
to
ensure
safety
and
optimality,
etc,
etc?
So
these
are
some
of
the
newer
problems
that
we
are
working
on
in
the
academic
community
and
fortunately,
or
coincidentally,
we
are
finding
at
least
a
few
companies
stepping
up
and
saying.
We
want
to
try
this
because
we
really
see
this
as
an
issue,
perhaps
a
small
issue
at
this
point,
but
increasingly
a
very
big
potential
issue
for
us
going
forward.
H
So
let
me
stop
there.
I
think
I
have
used
up
more
or
less
the
time
which
was
given
to
me
happy
to
answer
any
questions
either
now
through
the
chat
or
even
offline.
Later
on
you
can
you
have
my
contact
details
here
on
the
screen?
Thank
you
very
much
for
your
attention.
D
Thanks
a
lot
professor,
it
was
really
very
rewinding
and
insightful
also
and
happy
to
note
that
also
from
sap
other
technology
also,
we
are
working
with
most
of
our
industry
members
and
on
these
directions
also,
so
it
was
really
very
insightful
any
questions
or
we
can
post
you
in
chat
also.
So
there
are
any
questions
we
can
quickly
ask.
D
Before
so,
in
the
interest
of
time,
I
would
request
everyone
to
post
it
on
the
chat
questions
for
professor,
so
our
next
speaker
is
sir.
If
you
can
kindly
unshare
your
screen,
I
have
done
that
yeah.
Thank
you,
okay.
So
our
next
topic
is
going
to
be
taken
by
mr
alok
khanna
who's.
The
executive
director
strategic
is
indian
oil
and
he's
going
to
speak
on
the
indian
oil
digital
journey.
So
will
you
be
sharing
your
screen
or
should
I
share
the.
D
D
D
A
A
D
D
A
A
D
E
H
A
J
B
B
Just
enter
their
mobile
number
and
just
say,
call
me,
and
it
will
directly
system
will
call
him
on
his
mobile
number
or
his
desk
number.
That's
what
she
said
which
one.
B
So
if
you
click
on
the
meeting,
if
you
go
to
the
there
is
a
three
dots
right:
more
actions.
If
you
go
there,
there.
D
D
B
D
D
D
We
are
hearing
mr
khan
about.
Is
he
there.
G
Sandeep,
I
think
that
is
when
you
join
it
ask
for
the
name,
so
that
is
where
he
is
calling
the
name.
Okay,.
D
E
F
F
D
F
But
thanks
for
calling
me
in
during
this
meeting
and
thanks
for
putting
allowing
me
to
put
isbn
perspective
now,
indian
oil
had
thought
about
its
ip
strategy
way
back
in
2017
and
we
had
in
our
internal
history
also
found
a
lot
of
gaps.
So
basically
it
is.
The
id
study
study
was
conducted
to
align
it
really
with
the
removal.
F
So
the
10
gold
moves
were
in
next-gen
operations,
so
improving
the
our
operating
locations
and
operating
processes,
the
supply
chain
management,
commercial
excellence,
smart
maintenance,
safe
and
productive
environment,
smart
knowledge
management,
smart
office,
customer
first
catholic
excellence
and
data
and
analytics
so
I'll.
Take
them
one
by
one.
But
these
are
the
10
areas
where
we
found
that
yes,
it
could
lead
us
to
a
better
position
in
future.
F
We
are
talking
about
improving
these
refinery
performance,
having
deploying
tools
for.
F
F
We
are
trying
to
find
out
the
use
of
artificial
intelligence
in
having
in
improving
our
supply
chain
because,
as
you
know,
indian
oil
is
having
11
refineries
exploding
all
over
the
country.
Some
are
custom,
some
are
inline
refinery
and
then
we
are
distributing
product
to
almost
200
operating
locations.
F
F
F
F
We
were
not
able
to
use
the
historian
data
so
effectively
because
of
the
absence
of
proper
platform
and
proper
tools,
so,
of
course,
iot
is
going
to
be
deployed,
but,
as
you
know,
in
refinery
the
already
there
are
a
lot
of
sensors.
A
lot
of
data
is
already
being
collected
that
only
real
time.
This
is
only
thing
is
it
was
in
a
black
box
called
historians.
F
F
F
Safety
is
there
a
paramount
importance
so
how
to
deploy
technologies?
As
I
told
you,
one
system
we
have
developed
for
safety
is
the
monitoring
of
environment
parameters
online.
But
here
we
are
connected
worker
app
all
these
technologies.
We
have
found
that,
yes,
they
could
be
useful
because
it's
a
huge
workforce
working
in
our
refinery,
our
own
operators
contractor
operators,
even
monitoring
of
contractor
workers
in
a
refinery
just
to
ensure
that
he
doesn't
roam
around
or
go
goes
away
from
his
designated
working
area.
F
So
the
newest
technology
can
help
us
a
lot
in
visiting
and
ensuring
that
the
workers
remain
within
their
last
places.
Similarly,
training
is
one
area
where
we
are
found
and
we
are
deploying
it.
We
start
with
the
safety
training
prayer
via
technology.
F
The
enterprise
content
management
system,
the
workflow
systems,
so
these
are
these-
are
tools
which
which
are
getting
deployed.
In
fact,
this
has
already
been
deployed.
Similarly,
start
a
smart
office,
again
digitizing
the
workflows
now
this
has
held
us
in
good
state
in
the
code
19
scenario
where
now
a
lot
of
the
approvals
are.
F
Robotic
process
automation,
we
have
identified
as
a
big
opportunity
in
our
company
and
we
have
this
class.
We
have
already
started
implementing
the
rpa
and
there
are
many
many
redundant
processes
which
can
be
brought
into
our
field
and
crucial
manpower,
useful
and
power
can
be
saved.
F
F
Replenishment
for
the
retail
outlets
and
our
big
consumers,
we
are
also
doing
a
study
on
retained
extension
of
the
existing
customers,
so
contact
predicting
there
that
the
customer
is
going
to
leave
is
a
big
area.
Of
course,
we
are
trying
to
use
ai
to
and
alert
our
field
officers,
which
customer
may
leave
us
and
better
go
and
contact
him.
F
We
have
implemented
a
lot
of
capital
projection.
Even
now
many
capitalists
are
going
and,
of
course,
there
could
be
some
hold
up,
but
it
will
again
catch
up
so
capex
management
end-to-end
project
management.
F
Even
once
the
project
is
completed,
the
talk
nowadays,
the
documents
are
not
handed
over
in
digital
form,
always
so
it
could
be
physical
handover,
but
these
are
crucial
documents
which
we
need
after
5
years,
10
years,
also
now
to
digitize
these
things
and
seamlessly
hand
over
some
projects
to
the
operating
people.
F
F
I
F
F
F
The
problem
was
that
the
research
octane
number
was
measured
only
once
a
day
in
there
and
the
for
the
deadlines
after
24
hours.
The
refinery
is
done
on
that
one
reading
of
business,
so
we
deployed
the
ai
base
terms
that,
whether
depending
on
because
it
depends
on
five
six
parameters,
which
are,
of
course,
the
temperature
flow
rate
and
other
parameters
or
the
pressure
of
the
vessel
vessels
in
the
unit.
F
What
could
be
wrong
and
based
on
that
algorithm,
we
tried
to
predict
the
law,
and
this
model
was
run
for
three
months
in
that
refinery
and
when
the
refinery
officer
said
yes,
your
model
is
predicting
the
iran
properly
and
we
can
depend
upon
your
prediction.
Also,
of
course,
lab
report
was
available
only
once
a
day,
based
on
that
one.
Now
further,
we
developed
the
model
how
to
optimize
the
yield
and
or
on
both
or
so
to
say,
as
the
operator
feeds
the
desirable
wrong
number
at
that
point
of
time,
because
the
the
dc
extension.
A
F
Because
previously
speakers
have
also
talked
about
using
ai
in
analytics,
and
I
am
really
confident
that
the
refineries
our
refineries
or
for
that
matter
even
our
pipeline,
they
are
already
having
a
host
of
data.
It's
up
to
us
to
use
that
data
and
obtain
the
best
results
again.
I'll
share
you
a
very
small
example:
indiana
is
running
almost
100
bottling,
lpg,
bottling
plants
and,
as
you
know,
you
are
all
consumed
from
lte
that
it
is
always
in
shortage.
F
Of
the
controlling
offer,
could
there
be
a
better
way
of
deploying
ai
in
predicting?
If
I
run
the
plant
on
sunday,
I
will
be
able
to
run
this
plant
on
monday,
tuesday.
Also,
it
should
not
so
happen
that
I
run
the
plant
on
sunday
and
monday
and
tuesday.
I
don't
have
enough
indents
to
execute
and
of
course,
there
are
other
parameters,
whether
bulk
will
be
available
or
not.
How
much
bulk
lpg
is?
Is
there
standing
at
this
plant?
F
F
Through
this
digital
transformation
project,
we
are
deploying
enterprise
wide
iot
platform,
as
I
said
that
already
we
are
having
host
of
data.
So,
first
of
all,
we
will
like
to
bring
that
data
to
a
central
platform
augment
that
data
with
the
iot
sensors
wherever
required,
and
then
the
central
platform.
This
will
be
used
to
do
his
case.
Specific
analytics.
F
F
The
custom
solutions
for
their
own
equipment,
like
for
generators
or
for
turbines,
etcetera,
etcetera.
It's
a
question
of
supplying
the
data.
So
once
you
have
the
platform
you
can
take
the
decision
of
outsourcing
some
analytics
developing
certain
analysis
in-house,
giving
it
to
outsourcing
it
to
your
vendor
to
develop
solutions
for
you
and
the
choice
is
open
thereafter.
F
We
are
also
as
a
part
of
the
json
transformation
project.
We
are
also
assisting
our
hardware
requirements.
What
are
going
to
be
our
audio
requirements,
the
kind
of
analysis
the
user
is
requiring
today
today
the
user
is
requiring
the
analysis
of
the
flow
of
the
flow
from
the
nozzle.
So
we
have
40
000
or
30
retail
outlets
and
the
each
little
outlet
having
four
to
five.
F
The
dispensing
unit
having
and
delivering
the
product
to
customers
in
their
university
and
the
users
desire
online
analysis
from
each
nugget
the
kind
of
data
flow.
What
we
are
talking
about
requires
augmentation
of
hardware
also,
so
that's
what
we
are
doing.
We.
F
Video
analytics
policy,
which
is
very
interesting
subject,
and
we
are
doing
some
cases
on
video
analytics.
Like
simple
cases,
one
thing
returns
with
x
number
of
cylinders
in
the
plant
there
are.
F
So
now
a
man
is
to
be
deployed
to
count
the
number
of
cylinders
to
count
whether
you
see
when
a
cylinder
is
damaged
or
not,
and
many
other
parameters
can
the
media
analyst
will
be
deployed
to
release
that
man.
So
these
are
the
kind
of
projects
we
are
working
on
then
similarly
very
interesting
case
of
material
goods.
F
Material,
multiple
courts
have
been
generated
wrongly
by
users
over
the
last
few
years,
and
now
we
are
having
huge
number
of
material
codes.
Can
we
develop
some
algorithm
to
analysis
to
identify
the
duplicate,
materials
and
block
them
again?
This
can't
be
a
manual
process.
This
has
to
be
automated
at
times
you
may
have
to
deploy
rca
also
in
this,
but
we
are
working
on
and.
F
I
A
F
F
Doing
trying
of
new
technologies
implementing
those
new
technologies
scaling
up
of
those
new
technologies.
So
this
is
going
to
be
a
in-house
expert
tool
which
we
are
developing,
which
will
be
not
only
showcasing
what
technology
is
limited,
but
we
will
also.
F
That
these
are
the
change
management
exercises.
We
are
doing
as
a
part
of
this
project.
Every
company
is
doing
it,
so
we
are
also
doing
it,
but
we
are
trying
to
popularize
it
among
the
masses
within
the
company.
Just
to
give
you
an
idea
we
had
developed,
we
had
done
a
hackathon
on
involving
young
officers
in
december
and
it
was
a
huge
success.
It
was
participated
by
50
people
who
came
with
their
own
set
of
business
problems.
F
F
F
D
F
F
The
main
objective
will
be
obtaining
some
financial
benefit
out
of
it,
so
the
final
target
is
that
whatever
digital
project
we
are
trying
to
bring
in,
it
must
bring
some
kind
of
benefit
to
the
company.
F
F
F
F
D
D
However,
giant
is
actually
not
easy,
so
I
think
we
have
tremendously
benefited
I
in
the
interest
of
time.
I
will
request
everyone
to
post
their
question
on
the
chat,
and
I
may
I
request
my
next
speaker,
sandeep
and
varun,
who
will
come.
We
have
heard
a
lot
about
analytics,
but
let
us
also
hear
what
should
be
the:
what
should
we
can
be
a
reliable
and
robust
architecture
which
can
support
these
visions?
So
what
do
you
sandeep
and
varun
kindly
keep
it
within
20
minutes?
So
you
start.
G
Okay,
thanks
sandeep
and
thanks
alok,
sir,
for
this
great
presentation
on
iocl
initiatives.
G
So
myself
and
varun
will
try
to
cover
this
part
of
the
how
we
can
enable
these
analytics
and
also
we'll
show
the
demo
on
this
thing.
So
so,
basically,
I
hope
everyone
can
see
my
screen.
G
K
G
Exiting
okay
great,
so
so,
basically,
what
we
are
planning
in
next
20
minutes
is
that
how
big
data
and
analytics
solutions
from
sap
can
help
in
the
oil
and
gas
industry,
based
on
our
experience
so
before
getting
into
the
solution.
G
Just
wanted
to
set
the
context
here
that
why
we
require
a
big
data
or
analytics
solution
today,
as
we
are
hearing
from
multiple
speakers
this
morning
about
it,
and
so
in
our
opinion,
what
the
current
pace
of
technology
advancements
has
the
profound
impact
on
enabling
how
oil
and
gas
companies
transform
themselves
right,
and
there
are
certain
lenses
we
identify-
or
we
understand
from
the
our
experience
like
minimizing
the
cost
of
production
or
like
improving
the
performance
of
the
matured
wells
or
maybe
minimizing
the
shutdown
time
or
consider,
and
there
are
a
lot
of
environmental
regulations
as
well.
G
So
these
are
the
some
of
the
challenges
which
is
shaping
this
particular
thing
about
the
oil
industry,
island,
gas
industry.
So
this
with
this
thing,
I
think
see
based
on
certain
surveys.
What
we
have
heard,
or
what
we
understood
is
that
like,
for
example,
in
the
data
management
space,
74
percent
of
the
enterprises
say
their
landscape
is
very
complex
and
we
also
heard
the
similar
details
from
the
iocl
just
now
that
they
are
working
on
the
digitalization.
G
G
They
are
not
able
to
get
the
data
which
can
be
used
for
the
information
for
the
decision
making.
Basically
and
another
thing,
what
we
heard
is
that
55
percent
of
the
organizations
rate
their
alignment
between
id
and
business
as
moderate
or
worst,
so
what
it
is
have
to
offer
in
today's
time
is
not
what
business
is
actually
looking
for,
because
business
is
looking
for
something
more.
So,
for
example,
we
are
working
with
couple
of
oil
and
gas
companies
also
where
we
have
done
the
digital
strategy.
G
We
have
found
that
in
the
operation
side,
people
are
looking
for
more
predictive
kind
of
use
cases
and
they
have
full
knowledge
on
this
thing.
However,
getting
the
data
at
one
place
or
accessing
those
data
points
is
a
challenge
and
similarly,
from
the
intelligent
technologies
perspective,
64
percent
of
the
organizations
are
struggling
to
use,
predictive,
analytics
or
machine
learning.
So
these
are
the
some
of
the
challenges
which
we
have
identified
from
the
various
surveys
we
have,
or
the
studies
done.
G
Similarly,
when
we
talk
about
the
missing
link,
for
example,
so
what
happen
because
of
the
complex
landscape,
when
I
am
trying
to
collate
the
data
also,
it
is
difficult
to
bring
because
of
the
variety
of
the
data
types
of
the
data
like
some
is
coming
from
the
transactional
system.
Some
is
coming
from
the
sensors.
G
You
are
able
to
provide
that
data
point
or
provide
that
information
so
and
then
tools
and
technologies
are
also
like
there
are
specialized
technology
for
the
specialized
kind
of
requirement.
So
this
is
another
limitation
which
we
have
heard
so
and
based
on
these
complexities
based
on
those
limitations.
This
is
how
it
becomes.
You
have
data
in
multiple.
G
You
end
up
storing
data
at
multiple
places,
be
it
multiple
clouds,
be
it
multiple
systems
also-
and
this
has
become
this-
has
this
brings
the
complexity
in
our
system,
basically,
and
also
it
poses
a
risk
because
most
multiple
times
what
we
have
seen
based
on
our
experience
is
that
people
are
taking
the
dump
of
this
data,
putting
it
in
some
excel
or
in
some
other
tool
and
then
doing
some
kind
of
processing
on
this
and
then
producing
the
result.
G
In
that
case,
you
are
not
keeping
the
track
of
the
data
where
all
it
is
going,
so
so
that
is
also
the
from
the
legal
or
substantial
risk
from
the
data
security
perspective
also
there
so
considering
those
kind
of
challenges.
What
sap
is
reason
from
the
data
perspective
is
like
this.
So
already,
you
are
aware
about
the
top
layer
where
we
are
talking
about
the
digital
core,
like
employee
experience,
point
of
view,
various
applications.
G
Basically,
however,
from
the
middle
layer
perspective,
if
we
talk
about
so
on
the
middle
layer
perspective,
that
is
where
sap's
vision
is
that
you,
you
can
have
your
data
in
various
places,
not
a
problem.
However,
from
the
intelligent
technologies
point
of
view
in
sap,
what
we
can
try-
and
we
can
help-
is
that
bring
great
benefits
like
such
as
productivity
or
fa
efficiency
gains,
or
maybe
some
innovative
new
business
models.
Also,
we
can
bring
that
right
by
putting
the
models
and
new
revenue
streams
as
well
so,
for
example,
advanced
analytics.
G
It
can
help
get
real-time
visibility
into
your
operations,
customer
feedback,
maybe
or
changing
environment,
so
that
they
can
simulate
the
impact
of
business
decisions
and
mitigate
the
risk
also.
Similarly,
if
we
talk
about
the
integrated
advance
analytics
capabilities,
including
the
situational
awareness
into
applications,
enable
oil
and
gas
com
operators
to
analyze
all
type
of
data,
on
the
same
line
like
on
the
like
improving
the
decision
making
at
all
levels.
So
this
is
how
we
are
saps
vision
on
the
on
the
data
point
of
view
and
for
fulfilling
this
vision.
G
There
are
certain
guidelines
principles
we
have
defined
from
the,
for
example,
from
the
id
perspective,
we
are
talking
about
the
reducing
the
latency
or
redundancy.
So
what
that
does
that
mean?
Is
that
what
we
are
saying
is
that
you,
for,
if
you
want
to
use
certain
data
for
a
certain
part,
a
particular
purpose.
You
need
not
to
bring
that
data
into
a
particular
platform
to
make
it
a
single
source
of
truth.
You
can.
The
data
can
remain
wherever
it
is.
G
However,
when
I
require
it
for
the
processing
on
the
fly
on
the
on
demand,
it
should
be
available,
so
that
is
so.
We
are
trying
to
reduce
the
redundancy.
Similarly,
the
agile
iterations
see,
even
if
I
have
to
do
a
multiple
time
of
operations
or
multiple
iteration
in
coming
up
with
models
or
something
I
don't
have
to
create
multiple
data
copy
for
that,
for
example,
similarly,
like
minimizing
the
business
disruption,
so
what
we
are
saying
is
that
if
any
customer
or
our
business
users
are
coming
to
us,
our
ability
should
not
be
like
that.
G
Okay,
we
cannot
provide
you
this
data
at
this
point
of
time
because
of
certain
limitation
or
something
so
we
are
making
a
reusable
content
as
well,
at
the
same
time,
from
the
business
priority
perspective,
for
example,
enterprise-wise
visibility,
so
be
it.
I
have
a
digital
oil
field
system,
be
it.
I
have
sensors
on
my
machines,
be
it
a
transactional
system,
be
it
my
fuel
management
system,
maybe
or
maybe
the
power
consumption
systems.
G
I
can
see
all
this
data
at
one
place
in
real
time
from
the
from
maybe
in
the
dashboard,
or
maybe
I
can
use
this
data
for
the
protective
kind
of
use
cases
like
demand,
forecasting
or
maybe
predictive
maintenance
kind
of
thing,
similarly
time
to
value.
So
this
is
very
important.
Whenever
I
require
this
data,
it
should
be
available
for
me,
so
these
are
certain
design
principles
which
we
follow
and
basis
that
this
is
the
high
level
architecture
which
we
have
come
up
from
the
especially
from
the
oil
and
gas
perspective.
G
So,
for
example,
if
I
have
to
read
directly
from
the
iot
sensors,
I
can
use
either
my
iot
edge
device.
This
thing-
or
maybe
I
can
use
kafka
streaming
for
that
purpose
and
or
if
suppose
we
are
talking
about
the
historian,
like
maybe
ge,
professory
or
maybe
osi
pi.
We
can
directly
read
this
data
into
hana
from
this
plant
connector
perspective,
so
any
opc
compliant
historian.
G
We
can
connect
as
well
so
so,
basically,
one
important
factor
here
which
is
enabling
this
all
these
various
sources
is
something
called
sap
data
intelligence
in
the
middle,
so
what
it
provides.
It
is
basically
providing
the
enterprise
information
management
ability
to
surface
data,
govern
data,
examine
data
from
most
resources
with
enterprise
ai,
which
allows
us
to
learn
from
these
data
and
drive
basis
value.
So,
for
example,
in
in
this
particular
case.
G
If
I
take
an
example
of
predictive
maintenance,
for
example,
so
I'm
getting
data
from
the
sensors
using
my
iot
edge
devices
now
one
way
of
looking
at
it
is
that
I
can
get
this
data
store
it
first
in
the
data
lake
or
maybe
in
the
data
warehouse,
bring
my
essay,
because
I
am
talking
about
the
sensors
data
which
is
coming
from
my
machines.
I
also
need
to
get
the
asset
data
from
my
ecc
or
s4hana
system,
because
whatever
maintenance
has
happened
from
the
or
what
is
this
asset?
G
All
that
asset
information
maintenance
information
records
are
there
in
my
ecg
system.
So
I
have
to
bring
all
these
data
into
one
place,
maybe
in
the
data
lake
or
in
the
data
storage,
and
then
I
can
apply
certain
algorithms
on
on
top
of
it.
However,
there
is
another
way
of
doing
it
in
real
time
as
well.
So,
for
example,
I
can
deploy
or
create
my
my
model
in
python.
For
example,
I
can
deploy
and
data
intelligence,
so
I'm
getting
reading
this
data
from
sensors
in
real
time.
G
I'm
reading
this
data
from
ecc
or
s4rana
in
real
time.
I
am
creating
a
pipeline
in
the
data
intelligence
and
then
applying
applying
these
models
on
top
of
it.
So
and
then
I
can
consume
the
output
of
these
models
in
my
visual
at
the
visualization
layer
in
real
time,
without
necessarily
storing
this
data
in
the
data
warehouse.
So
that
is
one
of
the
capability
which
data
intelligence
provides.
So,
having
said
that
varied
sources,
I
ca
one
solution
can
take
care
of
processing
of
multiple
kind
of
data
points.
G
Secondly,
when
I
am
talking
about
the
varied
sources,
it
also
become
very
important
that
I
should
have
all
the
metadata
captures
of
from
the
source
system
so
that
tomorrow,
if
any
issues
are
there,
if
any
data
lineage
issues
are
there,
I
should
be
able
to
refer
it
back.
So
that
is
another
capability
which
this
data
intelligence
tool
provides
me.
So
so,
basically,
I
can
have
data
discovery
and
pipelining
monitoring,
as
well
as
business,
catalog
kind
of
thing
in
one
place
in
data
intelligence
and
also
suppose
from
the
reporting
and
analytics
point
of
view.
G
I
need
this
data
in
real
time,
so
I
can
use
this
smart
data
integration,
for
example,
to
read
data
from
s4hana
and
put
it
in
the
real
time
in
the
data
warehouse,
so
that
same
can
be
used
for
my
kpi
dashboard
or
visualization,
or
maybe
for
the
predictive
purpose
as
well
or
similarly
from
these
systems
homegrown
systems
as
well.
I
can
read
in
real
time
and
put
it
in
the
data
warehouse.
Now
when
I
come
to
the
storage
layer.
In
this
case
we
are
talking
about
the
bw
for
hana
plus
hana.
G
So
so
that
is
why
it
is
very
important
and
then
secondly,
from
the
hana
perspective,
I
can
do
the
sql
based
operations
also-
and
I
don't
have
to
write
any
a
base,
complex
program
to
get
my
data
or
get
the
queries
done
so
so
this
is
a
storage
layer
we
are
talking
about,
then
comes
the
predictive
layer.
So
one
more
important
thing
about
before
I
talk
about
the
predictive
layer
on
the
data
intelligence
side
is
that
we
have
basically
this
data.
Intelligence
is
a
very
open
tool.
G
We
are
not
restricting
anything
so,
for
example,
while
data
intelligence
provide
you
the
auto
ml
capability
by
default,
it
hana
pal
library
is
also
there.
However,
at
the
same
time,
you
can
also
bring
your
own
model
and
deploy
at
the
processing
layer
in
the
data
intelligence
itself.
So,
for
example,
I
have
a
python
model.
I
can
just
bring
that
model
and
deploy
here
in
the
data
intelligence
and
execute
it
similarly
from
the
r
as
well,
and
you
can
bring
any
other
statistical
tool
capability.
G
Also
here,
jupiter
notebook
is
already
part
of
the
data
intelligence
as
a
runtime
environment,
where
you
can
write
your
models.
Also,
you
can
do
the
here.
You
can
also
do
in
data
intelligence,
your
model
versioning
as
well
so
model,
so
you
can
maintain
the
version
of
the
models
as
well
in
data
intelligence,
and
so
so.
G
This
is
what
we
are
talking
about
from
the
predictive
layer
perspective,
any
predictive
library
you
can
bring
and
execute
or
maintain
in
your
data
in
this
particular
architecture,
without
any
limitation,
and
then
from
the
point
of
view
of
the
visualization
and
bi
layer,
we
are
talking
about
sap,
analytics
cloud
or
business
object,
bi,
where
you
can
so,
as
you
know
that
from
the
sap
analytics
cloud
point
of
view,
we
are
talking
about
bi
planning
and
predictive
all
three
in
one
tool
so
that
when
I'm
seeing
my
business
bi
report,
I
can
also
run
some
predictive
algorithm
and
find
the
inside
from
the
system
on
on
the
kpi,
which
I
am
use
seeing
right
now,
or
I
can
also
when
I
am
talking
about
the
financial
for
example-
or
maybe
I
am
talking
about
the
production,
I
can
also
see
the
planning
or
do
the
planning
at
the
same
time
in
the
same
tool
without
again
going
to
the
multiple
tools.
G
G
So
from
the
unstructured
data
perspective,
you
can
go
for
any
hadoop
kind
of
system
and
it
could
be
any
hyper
scalar
hadoop
like
from
ads
aws
or
maybe
azure,
or
maybe
some
open
source
hadoop.
You
can
download
and
install
and
configure,
and
you
can
directly
connect
this
hadoop
with
the
data
intelligence
for
from
the
processing
perspective
as
well.
So
if
you
have
any
question
on
this
thing,
please
let
me
know.
G
L
L
Yeah,
I
am
just
re-sharing,
can
you
see
it
now?
Yes,
okay,
so
maybe
I'll
take
a
quick
couple
of
minutes.
You
know
explaining
a
scenario
and
then
showcasing
it
in
action.
So,
as
sandy
pointed
out,
let
us
look
at
two
scenarios,
one
from
the
data
intelligence
for
my
artificial
intelligence,
workbench
and
then
talking
about
the
analytics
layer.
So
when
we
talk
about
the
you
know:
artificial
intelligence
layer,
there
are
multiple
problems
right,
one
is
connecting
the
data.
One
is
the
ability
to
look
at
data
managing
it
normalizing
it
pre-processing.
L
It
look
at
missing
values.
So
so
a
lot
of
speakers
already
highlighted
the
the
ability
to
have
usable
data
and
in
case
of
not
having
the
right
data,
how
do
I
manage
the
metadata
right?
So
a
lot
of
you
know
capabilities
to
manage
the
data
and
then
also
the
ability
to
manage
the
predictive
model,
create
it
manage
the
entire
life
cycle.
So
typically,
this
happens
across
silos
and
there
is
no
integrated
machine
learning
framework
to
end
to
end
manage
an
artificial
intelligence
project
within
a
typical.
L
You
know
company
scenario,
and
that
is
where
you
know
the
tool
comes
in.
So
I
just
give
you
an
example.
So,
for
example,
you
know
from
an
asset
right
taking
the
example
that
sandeep
talked
about
a
predictive
maintenance.
The
sensors
are
giving
me,
let's
say
a
pressure
and
temperature
reading
on
on
on
a
near
real
time
basis.
I
have
data
from
my
remote
asset
and
its
maintenance
schedules
and
stuff
like
that.
Coming
in,
I
have
my
erp
data,
where
I
have
the
master
data
of
the
asset
when
it
was
last
service.
L
Who
was
the
service
order?
I
mean
who
was
the
service
agent
and
stuff
like
that,
and
then
I
also
have
ir
cameras.
You
know
more
of
a
video
analytics
and
a
big
data
or
image
analytic
coming
from
drone
from
my
asset,
and
I
am
able
to
correlate
or
merge
all
the
data
as
a
metadata
for
my
modeling
and
from
there
I
am
trying
to
do
a
predictive
quality
or
a
predictive
maintenance
from
there
right.
L
So
typically,
you
know
when,
when
when
something
that
needs
to
be
done,
like
that
the
pain
points
of
somebody
who
wants
to
do
this
right,
a
data
scientist
or
a
data
analyst
has
a
lot
of
manual
effort
to
bring
all
of
this
data.
The
ability
to
deploy
manage
is
always
a
problem
right
now.
How
do
we
solve
this?
So
if
you
just
look
at
this
solution,
this
is
sap
data
intelligence.
Now
it
has
the
ability
to
connect
to
multiple
data
sources
like
sandeep
talked
about
it
can
be
sap,
it
can
be
a
hadoop
database.
L
It
can
be
a
standalone
csv
right.
I
can
automatically
bring
all
the
data
here.
The
system
has
ml
for
me
to
predict.
You
know
how
the
data
is
coming
in
and
if
I
want
to
encode
the
data
in
a
particular
way,
a
string
to
decimal
or
handle
a
missing
value,
I
can
do
that.
As
you
can
see,
this
is
a
send.
L
You
know
data
from
a
sensor
which
is
basically
of
a
particular
windmill,
so
I
am
getting
the
data
and
I
am
getting
you
know,
data
in
a
near
real
time
stream
and
I
am
encoding
it
for
the
predictive
analytics
the
way
I
want
the
moment
I
encode
all
the
data
that
is
coming
in.
I
can
straight
away
go
into
visualization
and
I
can
use
if
you
look
at
here,
I'm
using
a
jupiter
notebooks
based
based
on
python.
L
So
you
know
it
is
very
much
completely
supported
for
us
to
allow
you
know
the
open
source
to
work
with
this
tool.
I
can
use
an
open
source
code
here.
I
can
visualize
the
sensor
data
coming
in
real
time
and
then
I
can
build
a
pipeline
here
right.
So
this
is
the
most
critical
part.
How
do
I
visually
manage
data
sitting
in
various
sources?
Right?
My
big
data,
my
video
data,
my
sensor
data,
my
unstructured
data
and
my
erp
data.
How
do
I
build
pipelines
right?
L
If
you
look
at
on
the
left
side,
I
have
multiple
pipelines
available.
How
do
I
use
these
pipelines,
for
example,
a
kafka
or
an
open
source
or
an
mqtt
for
iot,
and
you
know
an
email
operator
for
getting
any
workforce
related
emails,
and
I
also
have
the
ability
to
bring
in
any
other
data
and
merge
it
together
in
a
very
visual
manner
and
look
at
whether
the
connection
is
active
and
do
all
of
my
machine
learning
work
here
end
to
end
right.
I
can.
L
Also,
so
the
real
you
know,
prowess
of
this
solution
is,
it
is
completely
you
know,
flexible,
to
use
open
source,
ai
ml
algorithm,
over
and
above
the
proprietary
sap,
algorithms
that
are
provided,
but,
more
importantly,
right
from
data
identification
to
data
management,
to
model
creation
model
retraining
and
also
sending
the
model
for
production
and
maintaining
that
it
is
always
performing
at
its
optimum
end
to
end
can
be
done
on
this
platform
right
today.
There
is
no
platform,
typically
anywhere,
that
we
see
where
ml
projects
are
done,
where
there
is
end-to-end
model
management
capability.
L
So
that
is
where
sap
data
intelligence
comes
into
picture
right.
This
is
just
an
example
of
the
result
right
it
can
be
a
mobile
application
where
I
can
take
a
picture
of
the
asset.
You
know
the
image
classification,
deep
learning
algorithms
can
be
deployed
here.
Based
on
that,
I
can
identify
the
asset
from
sap
erp.
I
can
look
at
when
it
was
last.
You
know
serviced
what
is
the
remaining
useful
life,
whether
it
is
working
not
working?
Well,
I
can
take
a
survey
from
within
this.
L
It
can
go
back
and
help
my
predictive
model
perform
better.
So
if
you
look
at
it,
no
matter
where
my
data
is
lying
from
a
holistical
data
management-
and
you
know
my
machine
learning
framework-
the
tool
allows
me
to
manage
this
completely
end-to-end
right.
So
this
is
just
you
know
on
the
data
modeling
and
the
predictive
piece.
Now
you
know,
assuming
that
the
data
has
come,
how
do
I
visualize
the
data
right?
L
So
sandeep
talked
about
the
stack
right
so
now
the
stack
was
talking
about
data
coming
from
various
sources,
encoding
it
for
analytics
and
using
the
visualization
layer.
Not
just
to
do
you
know,
post-mortem
analytics
of
what
happened
where
it
happened,
but
also
to
pick
a
needle
in
a
haystack
scenario,
ability
to
simulate
or
do
a
what,
if
ability,
to
do
a
predictive
based
on
historic
data
all
within
the
dashboard
itself
right
so,
for
example,
here
I
have
my
financial
overview,
my
top
five
projects,
my
gross
margins,
my
upstream
volume,
so
on
and
so
forth.
L
So,
for
example,
if
I
go
into
a
particular
platform,
I
have
the
ability
to
you
know:
drill
up,
drill
down,
and
things
like
this,
and
here
is
an
interesting
part
right.
So
I
have
my
various
business
areas.
I
have
my
p
and
l
statement
that
is
getting
populated
from
the
backend
system.
Now
I
can
go
and
say,
for
example,
because
of
you
know
certain
market
scenarios,
my
change
in
stock,
the
stock
is
reducing
because
input
is
coming,
our
input
is
not
coming,
but
outflow
is
there,
let's
say
in
terms
of
my
crude
or
whatever?
L
So
now,
if
I
say
that
this
is
going
to
go
down
by
15
my
stock,
what
impact
does
it
have?
Based
on
my
historic
data
in
my
earnings
expenses,
net
profit
margin.
I
can
do
that
simulation
as
a
business
user
on
the
fly
and
the
system
has
the
ml
capability
in
the
back
end
to
go
and
recalculate
and
tell
me
once
this
change
happens.
How
does
it
impact
every
single
one
of
my
kpis
right?
So
as
a
business
user?
It
is.
L
It
is
not
just
giving
me
the
post
bottom
of
what
has
happened,
but
also
gives
me
the
capability
to
do.
You
know
a
deep
dive
analysis
of
why
something
happened
and
what
happened
right.
So
this
is
from
a
financial
standpoint.
L
So,
given
you
know
the
time
I
will
just
you
know,
touch
up
on
a
few
topics,
but
if
you
look
at
here,
this
is
ultimately
having
all
the
data
that
I
might
possibly
need
from
a
decision
making
standpoint
right
upstream
production,
downstream
portfolio,
performance,
refining
margin,
shipments
capital
projects,
environment,
health
and
safety,
risk
assessments,
human
resource
related
information
and
contingent
worker
statement
of
work.
L
So
if
you
look
at
it
is
basically
a
collection
of
my
entire
business
process,
which
is,
you
know,
put
up
like
an
agenda
for
me
and
I
can
click
and
go
to
any
of
these
analytics
and
get
you
know
all
of
these
detailed
insights
right.
So
here
I
have
direct
labor
input
costs
and
all
of
that
again,
as
I
said,
this
can
allow
a
lot
of
what,
if
simulation
features.
So,
for
example,
I
say
you
know,
we
all
know
that
due
to
covet
situation,
the
labor
is
a
challenge.
L
So
let's
say
in
my
particular,
you
know:
field
in
the
gulf
of
mexico.
Labor
is
a
challenge,
for
example
in
the
month
of
july,
not
so
much
in
the
month
of
september.
So
let's
say
the
labor
is
going
to
go
down
by
26
percent.
What
impact
does
it
have
on
my
operating
margins
and
my
overall
total
cost
of
production
by
barrier?
I
will
I
can
do
this
simulation
and
based
on
that.
The
entire
cost
is
cal
recalculated
in
the
backend
engine
and
it
comes
and
tells
me
what
impact.
What
does
you
know?
L
Reduction
of
labor
mean
for
me
in
terms
of
dollar
value
and
in
terms
of
my
production
capacity
output
right.
So
that
is
where
the
simulation
capability,
and
you
know
the
what
if
capability,
really
adds
value
to
the
dashboards
that
are
there
right.
So,
similarly,
if
I
go
into,
for
example,
refining
margin,
you
can
look
at
some
of
the
kpis
here
and,
and
the
important
point
to
notice
here
we
are.
L
This
is
something
that
you
know
as
kpis
sap
provides
out
of
the
box,
because
we
do
understand
that
in
an
oil
and
gas
industry,
it
is
not
easy
to
identify
all
of
the
kpis
across
all
of
the
lobes
identifying
where
the
data
needs
to
come
from.
So
what
sap
has
done
as
one
of
the
thought
leaders
in
the
oil
and
gas
industry
is
that
we
provide
some
out
of
the
box
content
called
sap
analytical
cloud
content
which
are
provided.
You
know
for
the
analysis.
L
L
You
know
that
kind
of
an
analysis,
so
any
needle
in
a
haystack
kind
of
a
scenario
which
usually
used
to
take
a
lot
of
time
and
a
very
deep
down
data
analyst
can
now
be
driven
by
the
system
using
machine
learning
and
help
the
business
user.
Look
at
you
know
very
in-depth
scenarios
and
decision
making
on
the
fly
right.
It
has
all
of
the
natural
language
and
those
capabilities
for
us.
Similarly,
this
is
my
supply
chain
performance
that
is
coming
in.
I
have
my
carrier
performance
dashboard.
Sorry.
A
L
So
basically,
I'm
done
so.
These
are
you
know
some
of
the
kpis
that
we
can
filter
and
look
at
and
and
the
point
to
notice.
These
are
all
completely
you
know,
not
just
sap
data,
but
data
from
multiple
sources
merged
with
the
sap
data
and
brought
to
the
you
know,
data
warehousing
layer
from
where
I
am
able
to.
L
D
I
think
we
have
this,
we
end
the
session,
so
you
can
give
the
thank
note
and
then
we
can
have
a
quick
pulse
check
in
the
ventimeter.
A
A
H
A
A
Here,
thank
you
very
much
everyone
for
joining
today's
session.
I'm
sure
you
would
have
found
value
in
in
the
sessions
as
well
as
the
speakers
that
that
spoke
today.
I
have
actually
enjoyed
enormously,
so
I
hope
you
all
have
too.
You
know.
I
just
want
to
extend
sincere
thanks
to
sap
team,
specifically
mukesh,
sandeep
and
money,
for
putting
this
whole
agenda
and
content
together
and
to
everyone.
We
look
forward
to
your
participation
in
the
next
sig
or
next
oil
and
gas
sig.
Also
you
have
a
great
weekend
and
be
safe.
D
Okay,
so
before
we
wrap
up,
can
we
again
go
to
the
mentimeter
screen
so
saran?
We
can
just
finish
the
four
questions
which
are
related
to
the
session,
how
it
went
what
things
needs
to
change
and
we
can
have
a
quick
wrap
up
on
that
so
over
to
everyone,
please
go
to
www.menti.com
enter
the
code,
836174.
D
D
A
A
D
D
So
this
is
something
you
can
just
type
in
what
topics
would
you
like
to
have
in
the
next
session
short
which,
if
something
which
you
have
heard
this
time
and
next
time,
you
want
to
hear
more
on
it,
so
you
can
highlight
that
and
on
any
new
topic
you
can
just
put
it.
D
D
Digital
strategy
advance
analytics
more
on
customer
side,
analytics:
okay,
hml
ito
t
integration
in
refinery
using
mai;
yes,
that's
a
very
good
use
case,
predictive
analytics
session
by
customers,
and
maybe
oil
and
gas
offerings
something
on
how
large
organizations
are
transforming
using
sap
project
system
for
monitoring
and
managing
capex.
Yes,
we
have
a
shutdown
and
around
topic
also,
which
we
mean
itot
integration,
analytics
iot,
digital
transformation
and
crm
topic.
Yes,
that's
this
time
we
were
more
on
assets
and
less
on
customer.
D
D
Yeah,
so
we
have
once
a
quarter
18
respondents
four
for
once
in
six
and
one
prefer
in
person
sessions
only
that's
a
very
good
revelation
and
there's
a
last
question
right,
siren.
D
D
D
D
So
yeah
reduce
the
number
of
topics
and
increase
content
per
topic
is
a
clear
answer
and
definitely
will
keep
a
note
of
it.
So
with
that
yeah
the
pulse
survey
ends
anything
from
manosa.
You
want
to
address
anything.
A
A
Thank
you
recipe.
Thank
you
mukesh,
because
you
know
thanks
for
your
efforts
to
get
this
first
virtual
meeting
of
the
sig,
I
think
it's
more
effective
and
productive.
We
take
the
feedback
from
these
questions.
Answers
provided
and
try
to
do
this
kind
of
activity
once
in
a
quarter,
and
maybe
more
customer
oriented
where
the
customers
share
their
success
stories
will
have
more.
So
once
again,
thank
you
very
much
be
safe,
stay
at
home.
Thank
you
very
much.
D
Sir
sir,
you
can,
can
you
just
switch
off
the
recording
okay
I'll,
do.