►
From YouTube: Maturity Model for Microsoft 365 - March 2023
Description
Recorded:
• March 21, 2023
Topic:
Cognitive Business Maturity Model Competency
The role of AI/ML etc. in business
Agenda:
• Practitioner & Maturity Model Overview
• Artifacts and Updates
• Purpose
• How to get more involved
• Shout-out & Picture Time (Together Mode!)
Hosts:
• Marc Anderson, Sympraxis Consulting | @sympmarc
• Emily Mancini, Sympraxis Consulting | @eemancini
• Simon Doy, iThink 365 | @simondoy
• Simon Hudson, Novia Works | @simonjhudson
• Sharon Weaver, Smarter Consulting | @sharoneweaver
Resources:
• Maturity Model for Microsoft 365 - https://learn.microsoft.com/en-us/mic...
• Interested on joining these calls - Download the invite at https://aka.ms/mm4m365/invite
A
So
welcome
everybody
who's
here
and
for
those
that
haven't
arrived,
I'm
sure
you'll
catch
up,
so
we've
slightly
changed
what
we
were
going
to
do
today
we
were
going
to
have
a
an
open
conversation
about
the
the
about
whether
we
should
build
a
model:
A
maturity
model,
competency
for
AI,
but
we've
gone
and
built
one.
So
so
we've
got
it
anyway.
A
Now,
so
it's
going
to
be
a
little
bit
more
looking
at
what
we've
built,
but
I
still
really
need
the
feedback,
so
yeah
so
feel
free
to
absolutely
jump
in
at
any
point
as
we
go
through
this.
So
some
of
the
questions
we've
got
in
the
slides
are
still
extremely
valid
because
you
know
when
I
started.
Writing
this
competency
two
weeks
ago.
Some
of
the
things
then
are
different
to
what
they
are
now,
because
that's
how
fast
it's
moving
anyway,
so
don't
jump
straight
into
the
slides.
A
A
We're
going
to
do
the
standard
grabbing
of
a
picture.
It
is
at
this
point.
I
always
wish
I'd
brushed
my
hair
beforehand
and
we're
going
to
invite
people
to
contribute
or
getting
get
involved
in
many
ways
and
then
we'll
get
on
to
the
main
event
which
I
I
apologize
in
advance,
but
will
be
mostly
me
talking
today
since
I'm,
the
only
one
of
the
group
that
really
knows
this.
This
new
competency
inside
out
since
I
literally
finished
it
two
hours
ago.
A
A
It's
a
clickable
link
because
we're
doing
it
in
PowerPoint
live
just
to
watch
out,
though
we've
got
MVP
Summit
coming
up
and
we
are
going
to
have
to
shift
the
date
of
one
of
the
or
the
next
session
to
accommodate
that.
So
we'll
give
you
a
bit
more
of
an
update
on
that
later,
but
this
invitation
Series
has
not
got
that
change
built
in.
A
Right
and
so
I
will
have
to
excuse
out
of
the
way
it
doesn't
work
anyway.
It'll
work
so
always
put
the
purpose
up.
On
the
left
hand,
side
you
can
have
a
read
of
that.
What's
happened
is
that
the
the
set
of
artifacts
on
the
right
has
been
reorganized,
because
we
have
have
now
an
exciting
new
cognitive
business
artifact,
which
is
finished
and
I,
dare
say,
We'll
Be
Live
by
the
end
of
the
week,
when
Mark
has
done
his
magic
and
then
marks
often
pushed
it
through
the
publication
process.
A
A
Right
just
to
remind
us
all.
The
maturity
model
is
a
five-step
process
and
from
the
initial
level,
which
is
very
much
that
sort
of
relatively
chaotic
chaotic,
not
very
thoughtful
ad
hoc
we're
just
trying
to
keep
the
lights
on
kind
of
approach.
Then
we
move
up
the
upper
levels
through
manage
into
defined,
which
we
really
consider
to
be
the
the
right
place
for
people
to
feel
secure
in
terms
of
running
a
business
or
organization,
and
then
we
go
up
from
there
into
optimizing
the
organization
through
an
intermediary
step
of
predictable.
A
All
of
this
stuff
and
more
detail,
besides,
is
in
the
the
maturity
model
overview
and
the
each
individual
maturity
model.
Competency
on
the
Microsoft
learn
site
right
now.
This
is
where
it
gets.
Scary
I
have
to
try
and
figure
out
how
we're
going
to
jump
out
of
out
of
PowerPoint
sharing
mode
into
the.
Let's
see,
everybody's
video
grab
the
screenshot
and
then
back
again.
So
let's
try
that
I.
A
A
A
A
A
I,
don't
know
back
in
the
frame
Cooper
that
was
completely
seamless
and
nobody
noticed
otherwise
right.
Okay,
so,
just
to
to
remind
you,
you
know
we
absolutely
invite
him
perhaps
almost
demand
contributions.
So
in
this
session
you
know
feel
free
to
stick
your
hand
up
we'll,
monitor
that
and
if
you've
got
live
questions
as
we're
talking,
please
do
so
do
use
the
chat,
we're
always
monitoring
that
and
we'll
try
and
either
capture
or
answer
the
questions
as
we
go
along
yeah.
A
A
E
A
Okay,
look:
we
don't
care
that
much!
It's
all
fine
right,
so
this
is
our
standard
overview
of
the
the
purpose
of
the
these
practitioner
sessions.
So
it's
a
safe
place.
Anyone
can
come
in
and
have
a
talk.
It's
a
way
of
us
getting
feedback
and
sharing
the
the
developments
around
the
the
the
maturity
model
review,
and
hopefully
you
if
you've
got
any
insights
around.
A
How
are
you
using
it
or
what
you'd
like
to
use
it
for
you
can
do
so
and
there's
a
bunch
of
links
there
to
other
ways
that
you
can
get
involved.
The
most
important
thing,
and
especially
going
to
reiterate
this
for
this
brand
new
competency
because
it
hasn't
been
through
a
review
process
so
far
is
when
we
get
it
published.
Please
use
the
feedback
process
within
the
the
Microsoft
learn
which
will
take
you
into
GitHub
and
raise
any
questions,
suggestions
or
changes
in
there.
A
It's
really
important
to
us
that
you
continue
to
get
involved
in
this
content
and
I
cannot
I'm
about
40..
Please
there
we
go
right
so
here
we
are
in
March
doing
the
AI
as
it
was
originally
built
now,
cognitive
business
next
month.
We're
looking
at
doing
a
a
session
on
Loop
and
Mark.
Is
it
next
month
that
we're
Shifting
the
data?
Is
it
a
month
after.
B
A
Right,
just
a
quick
call
out
to
those
people
that
came
along
last
time
and
again
just
to
give
you
a
couple
other
ways
that
you
can
get
involved
in
this
broader
Community
rather
than
just
in
the
practitioner
stuff.
So
that
slide
is
not
quite
correct.
Now
it
says
April
the
18th,
but
as
Mark
just
said,
it's
the
25th
next
time
around
and
we've
got
the
there
are
the
Marvels
Daryl
Webster
speaking
for
us
at
that
next.
A
Right
and
then
one
of
the
things
we
started
working
on
this
year
are
what
we're
calling
practical
scenarios
these
are
sort
of
at
the
next
level
down
from
the
competencies
and
where
we're
trying
to
so
I've
just
seen
the
picture
of
Kevin
on
the
right.
There.
That's
hilarious,
where
we're
trying
to
actually
give
more
Hands-On
advice
of
how
to
use
some
of
the
Technologies
and
use
those
to
move
you
up
to
the
competencies
and
again,
if
anyone's
got
any
views
on
those
or
would
like
to
go
and
write.
Some
that'd
be
horrific.
A
A
What
was
the
introduction
of
AI
and
machine
learning
and
related
Technologies
into
the
maturity
model?
Is
our
proposal?
Is
we're
going
to
call
it
cognitive
business
for
two
reasons?
One
is
that
ala
isn't
the
whole
story.
You
know
there
are
a
bunch
of
related
Technologies,
of
which
true
AI
is
only
one
of
them
and
also
as
we
thought
about
it,
because
the
maturity
models
are
really
around
business,
not
about
the
technology.
This
is
more
about.
How
do
we
use
thinking
in
business?
A
How
do
we
use
those
cognitive
processes
in
business,
and
so
we
came
up
with
the
term
cognitive
business
rather
than
Ai
and
machine
learning
and
other
kind
of
stuff
and
the
maturity
model.
So
that's
that
proposal
if
anyone
has
any
better
ideas
than
that
or
thinks
that
it's
a
rubbish
name
and
wants
to
push
back
absolutely.
Let
us
know
yeah
thanks
for
that.
B
A
B
Wanted
to
we
were
trying
to
come
up
with
something
that
was
generically
broad
enough,
that
additional
things
that
come
along
would
sort
of
slide
into
it.
If,
if
that
makes
sense,
you
know
the
the
AI
That's
The
Rage
today
or
was
six
months
ago,
is
probably
going
to
change
in
five
minutes.
We're
now
in
that
that
phase
I
think.
C
Yeah
and
also
it's
about
having
it,
you
know
targeted
at
the
business.
Isn't
it
and
you
know
sort
of
you
know
something
which
may
tweak
a
business
owner
or
senior
Management's
interest
in
it
as
well
and
I
like
the
idea
of
the
thinking
business
I.
Think
that's
really
really.
You
know
interesting,
but.
A
Yeah
to
have
taken
the
opportunity
to
sort
out
for
a
an
old-fashioned,
Wired
Mouse
that
I've
got
some
control
again.
Hopefully
there
we
go
right,
attend
this
session,
we're
originally
going
to
think
about
why
cognitive
business,
you
know
what
it
is,
and
the
role
of
AI
and
ml
in
the
maturity
model,
and
obviously
it's
quite
a
broad
topic
and
moving
very
quickly.
A
We
really
want
to
think
about
well,
how
does
this
AI
stuff,
or
these
thinking
Technologies
help
us
drive
organizational
success
and
model
the
kinds
of
problems
that
we
can
solve,
but
I'm
particularly
interested
in
the
the
relationship,
I
guess
between
the
cognitive
systems
and
the
cognitive
humans.
So
there's
a
bit
of
stuff
that
we're
thinking
about.
But
how
do
these
things
to
work
alongside
each
other?
A
And
if
any
of
you
saw
the
announcement
last
week
that
Microsoft
did
around
co-pilot
for
office
and
the
related
stuff,
they
had
quite
a
nice
phraseology
that
we
were
already
sort
of
playing
with.
So
it
was
kind
of
spooky,
but
they
talked
about
having
autopilot,
which
is
where
we're
kind
of
at
at
the
moment,
and
then
this
introduction
of
AI
that
acts
as
a
co-pilot.
It
works
with
people
rather
than
four
people
or
replacing
people,
and-
and
so
we've
been
spending
a
bit
of
time.
A
Also,
we've
talked
about
AI
in
the
maturity
model,
competencies
previously
and
typically
they
kick
in
at
level
400
and
we
undenired-
and
we
debated
for
some
time
about
whether
these
Technologies
deserved
a
competency
of
their
own
we've
got
some
precedence
for
this.
Things
like
Governors,
also
appear
in
many
many
of
the
competences,
but
now
have
a
competency
in
their
own
right.
A
There's
a
bunch
of
different
Technologies
and
more
are
emerging
all
the
time
again.
If
you
haven't
watched
the
the
announcements
from
Microsoft
last
week,
it's
worth
going
and
reading
the
blog
and
watching
the
Youtube
video
there's
a
lot
of
stuff,
and
some
of
it
was
unexpectedly
potent
and
well
developed,
but
we
already
are
using
AI
in
many
systems.
Some
of
them
are
relatively
invisible,
AI.
So
a
lot
of
your
threat,
detection
threat
prevention,
Technologies
built
into
Azure
and
Microsoft
365,
Rai,
driven
and
invisible.
You
don't
know
they're
there.
A
They
just
get
on
and
do
the
job
behind
the
scenes.
Then
there
are
some
which
you
have
to
set
up
like
syntax,
where
you've
got
to
do
a
bunch
of
work.
Training
up
on
your
content,
give
it
some
some
knowledge
to
work
on
and
the
way
it
goes.
And
then
there
are
those
ones
which
are
sort
of
very
front
and
center,
where
they
either
have
to
be
developed
by
you.
A
It
might
be
a
configuration
or
it
might
be,
a
coding
job,
but
there's
a
development
process
and
those
which
are
directly
interface
with
either
your
clients
or
with
your
staff.
So
we've
got
these
different
levels
of
AI
and
we
haven't
really
made
a
big
attempt
to
segregate
or
classify
those
out.
We've
looked
at
the
relationship
between
all
of
these
Technologies
and
the
way
that
business
works.
A
A
It's
really
important
for
us
to
observe
that
we're
early
in
the
stocking
curve.
You
know
we're
both
of
us
on
this
call
are
probably
going
to
be
in
the
enthusiasts
and
Visionaries
we
we
like
to
claim
they
were
pragmatic,
but
sometimes
our
enthusiasms
will
get
the
better
of
us.
We
need
to
understand
that
lots
of
other
people
are
going
to
be
lacking
on
that
and
we're
going
to
be
getting
into
the
pragmatists
and
those
that
come
there
after.
A
So
we
need
to
understand
that
and
that
there's
going
to
be
some,
you
know,
fear
uncertainty
and
doubt
terminal
use
again.
I
suspect
around
how
AI
interfaces
with
people
and
businesses,
and
also
important
to
note
the
technology
is
moving
so
fast
that
some
of
this
stuff
is
going
to
be,
have
to
be
Rewritten
or
were
replaced
in
the
not
distant
future
as
more
Revelations
come.
A
A
A
There's
lots
of
other
AI
stuff
out
there
already
and
some
of
those
are
invisible
and
some
of
those
are
are
visible.
So
those
of
us
in
the
coding
world
have
been
looking
at
co-pilot
for
GitHub
for
a
while
I'm,
not
a
developer,
but
I've
talked
to
quite
a
lot
of
developers
and
I
haven't
had
a
single
one,
go
it's
terrible,
they've
all
gone,
oh
I'm,
using
that
quite
a
bit
already
I
create
a
lot
of
content.
A
Microsoft
editor
I
I
considered
myself
to
be
top
five
percent
when
it
comes
to
use
of
English
language,
I've
learned
quite
a
lot
from
the
suggestions
that
Microsoft
editor
has
been
making
in
my
writing
style
and
my
grammatical
use.
So
again,
it's
trained
me
to
be
a
better
expert,
which
I
think
is
quite
impressive
and,
as
we've
just
said,
Bing
is
already
AI
powered
as
of
well
a
couple
of
weeks
ago.
Really
just
a
note
on
that.
As
far
as
I
can
tell
Bing
is
using
chat,
GPT
version
4.0.
A
A
I'm
not
going
to
read
this
stuff
out
but
yeah.
This
was
from
about
two
or
three
weeks
ago,
and
this
was
the
the
responses
that
we
were
getting
from
being
at
that
time.
Already
some
of
this
stuff
has
changed,
which
is
kind
of
nice,
but
when
I
was
talking
to
some
lay
people,
some
smart
lay
people
a
month
ago
about
the
emergence
of
these
chat
engines.
These
intelligent
chat
engines.
A
They
were
throwing
up
lots
of
criticisms
about.
How
do
I
know
it's
going
to
be
using
robust
sources
and
yeah,
and
you
know,
can
we
trust
it
and
all
these
kind
of
sensible
questions
and
by
the
time,
I
started
working
on
the
competency?
Some
of
those
art
questions
had
already
been
answered.
Yeah
they're,
starting
to
cite
their
sources.
You
can
click
through
from
any
of
the
statements
made
through
into
the
sources
as
it
as
it
runs
its
analysis.
You
can
see
the
chat.
A
The
Search
terms
is
using
to
pull
that
content
in
order
for
it.
To
summarize
that
so,
I
think
there's
a
lot
of
really
clever
stuff
going
on
behind
the
scenes,
which
is
around
the
governance
and
compliance
stuff.
Within
These
engines
and
that's
important.
We
talk
about
that
quite
a
lot
in
the
competency
right.
So
this
is
what
this
is
like.
A
So
they
may
be
using
things
like
editor
and
they
might
be
using
design
engine
within
PowerPoint
and
they
might
be
already
using
the
auto
tagging
that
things
like
Bing
and
Google
give
them
for
uploaded
images.
But
they
don't
know
that's
an
AI
and
they're,
not
that
interested
where
you
do
try
and
have
a
conversation
with
the
with
level
100
organizations
about
AI.
They
appreciate
away
all
the
express
Express,
fear.
Uncertainty
doubt
dismay
all
those
kinds
of
emotional
responses.
A
Most
staff
are
not
interested
or
not
aware,
but
some
people
are
saying
what,
if
we
used
AI
for
this
or
could
AI
solve
that
problem,
there
may
even
be
some
isolated
instances
of
people
experimenting
with
some
of
the
AI
services
to
see
if
it
could
add
value,
but
there's
no
real,
coherent,
there's
not
there's.
Just
no
is
no
coherent
strategy.
Senior
Management
either
are
uninterested
or
or
are
letting
it
run
its
own
course
for
now
and
not
putting
their
arms
around
it.
A
So
speculation,
perhaps
the
misconceptions
leading
to
you,
know
sort
of
negative
responses
and
any
experimentation
there
is
with
AI
is
disconnected,
but
there
might
still
be
some
value
at
that
level
by
the
time
we
get
to
level
300
and
I'd
be
surprised
if
there
are
many
organizations
at
300.
Yet,
given
how
new
this
technology
is
in
the
scheme
of
things,
then
level
300
they're
developing
their
own
awareness
of
cognitive
business.
This
is
the
business
of
using
cognitive
tools,
Ai
and
other
engines,
alongside
their
people
or
for
their
people.
A
A
Much
at
this
point
if
they
are
they're,
relatively
simplistic,
chatbot
type
applications,
any
kind
of
management
and
governance
at
this
point
is
almost
certainly
in
the
hands
of
I.T
is
not
considered
to
be
a
business,
critical
or
business
owned
system.
It's
it's
not
got
to
that
level
of
maturity.
Yet
so
there
may
be
limited
strategy,
but
it's
limited
the
oversight
that
might
be
required
mostly
isn't
there,
but
there
might
be
some
people
at
least
saying
you
know
every
chapter.
This
thing
has
not
got
bias
every
check.
A
This
thing
is
giving
us
some
reasonable
results,
but
they
may
not
be
measuring
what
a
reasonable
looks
like
at
this
point.
I
is
probably
not
treated
as
disruptive,
whether
that's
internally
disruptive
to
their
existing
processes
or
disruptive
in
the
marketplace
in
which
they
operate.
Although
there
may
be
people
beginning
to
worry
about
the
impact
that
their
competitors,
if
they
have
an
AI
driven
business,
would
have
on
them,
but
that's
about
as
far
as
it's
gone.
A
We
move
up
to
400
well,
from
the
model
point
of
view
we're
seeing
AIS
starting
to
appear
in
other
competencies,
with
one
of
those
efficiency
and
feedback
mechanisms
within
our
other
competences.
We've
seen
that
consistently
across
the
the
maturity
model
at
400,
you
start
to
see
three
things
really
emerging:
AIS
feedback
processes
and
the
emergence
of
the
organizational
culture
as
one
of
the
dominant
activities.
A
So
we
have
that
400
we're
also
seeing
we
would
expect
also
see
companies
that
have
got
an
active
strategy
around
Ai
and
that's
aligned
with
some
effective
management
processes.
There's
some
interesting
things
I'm
going
to
just
touch
on
again
in
a
moment
around
HR,
like
review,
we're
getting
into
people
understanding
what
the
ethics
and
other
responsible
use
of
AI
are
and
there's
a
whole
bunch
of
stuff
that
Microsoft
and
others
have
now
published
on
what
they
call
responsible
AI.
A
So
people
are
starting
to
look
at
their
AIS
and
going
does
this
actually
meet
our
internal
values?
Does
it
expose
us
to
external
sort
of
corporate
resource
corporate
responsibility
risks
by
the
way
that
we're
using
and
and
building
AIS
organization
at
level
400
are
starting
to
not
just
augment
or
enhance
their
processes?
They're
using
AI
Technologies
to
transform
their
processes?
That's
important,
and
they
are
thinking
about
what
is
the
relationship
between
the
people
and
the
Machines?
A
I,
would
guess
that
there
can't
be
more
than
a
handful
of
organizations
that
can
even
claim
to
have
level
500
cognitive
business
right
now,
but
some
some
will
be
looking
at
it
so
level
500.
We
think
that
you've
got
human
machine
teamwork.
It's
not
subservient
these
things
work
alongside
each
other
and
the
sentient
part
of
that
relationship.
The
people,
because
we're
nowhere
near
sentient
machines
right
now.
The
sentient
part
of
that
appreciates
the
capabilities
that
the
AI
gives
them.
That's
different.
A
The
other
thing
that
I
really
suddenly
had
a
a
an
Insight
moment
about
about
a
month
ago
was
again.
There
are
no
such
things
as
general
AIS
right
now,
artificial
general
intelligences,
but
one
of
the
transitions-
that's
happened
in
the
last
six
months
is
the
move
from
very
narrow
single
task,
basically
AIS
to
generalized
narrow
AIS
to
AIS.
That
can
do
many,
many
different
things,
so
one
AI
that
can
handle
many
different
classes.
A
problem
and
chat.
Gpt
is
an
example
of
that
yeah.
You
can
write
you
some
lyrics.
A
It
can
answer
a
complicate,
your
technical
questions.
It
can
pass
exams
for
you
and
all
sorts
of
other
things,
but
they
they're
up
into
the
hundreds
of
different
classes
of
tasks
that
they
can
do,
and
we
think
that
when
we
get
to
level
500,
these
kinds
of
generalized
AIS
will
be
able
to
do
hundreds
of
different
tasks
for
you
within
your
organization.
That's
not
hundreds
of
different
AIS.
That
is
a
handful
of
different
AIS
with
multiple
capabilities.
A
They
use
the
AIS
to
start
to
change
their
strategy
and
perhaps
even
to
develop
their
strategy.
The
business
processes
change
as
a
consequence.
The
AIS
are
pretty
much
ubiquitous
within
the
organization,
so
much
to
other
people
stop
noticing
that
it's
a
thing.
It's
a
piece
of
the
background
workplace
in
the
same
way
that
telephones
and
the
internet
are,
and
at
this
point
organizations
have
got
a
sophisticated
approach
to
that
responsible.
Ai.
You
know
they've
got
proper
management
stuff,
they
hold
their
AIS
to
account.
A
They
have
ethics
built
in
by
design
that
accountability
thing.
This
is
a
an
idea
that
we've
built
put
into
the
maturity
model,
but
I
confess
I
stole
from
a
colleague
at
capgemini
for
a
few
years
ago,
and
he
suggested
that
AIS
that
replaced
the
roles
of
people
or
fill
the
roles
of
people
should
be
treated
as
if
they're
people
they
should
have
line
management
structures.
They
should
be
put
through
performance
reviews.
A
You
know
at
least
analogous
ways
to
the
way
you're
doing
the
person
and
I've
held
on
to
that
idea.
I
think
it's
a
really
powerful
idea
of
not
treating
machines
as
machines,
but
treating
them
as
people,
because
they're
interacting
with
people
and
doing
things
that
people
would
do
and
and
if
we
do
it
right
and
give
our
AIS
a
good
ethical
base.
A
There's
the
promise
of
most
of
the
time,
then
being
better
ethics
than
people,
because
they
don't
get
hung
up
on
the
emotional
context
if
we
can
train
them
right,
there's
a
whole
philosophical
debate
about
that
that
I
might
come
to
another
day
right.
So
this
is
just
a
summary
of
some
of
that.
This
is
just
the
people
interface
at
the
low
levels.
People
do
it
at
the
lower,
slightly
lower
levels.
B
A
E
I
so
I
wanted
to
say
on
the
400
500
level
from
a
technical
perspective,
the
the
artificial
intelligence
and
behavioral
analytic
Tools
in
it.
Network
management
work
well
when
a
tenant
is
properly
licensed.
So
from
an
E5
perspective,
they
have
a
sort
of
Auto
remediation
Auto
resolution
when
certain
alerts
will
come
up
based
on
email
activity
or
SharePoint
or
OneDrive
and
sort
of
Insider
risk.
E
A
A
E
And
that's
where
the
tool
you
approve,
that
it
can
run
its
own
automated
response
and
then
you
get
the
notification
and
then
you
are
responsible
as
a
human
to
go
in
and
confirm
that
it's
a
false,
positive
or
a
false
negative.
And
then
you
can
reverse
the
actions
that
were
taken
or
then
you
know,
then
you
go
through
and
manage
the
response
on
your
own.
A
C
C
It
is
that,
and
so
is
that
what
level
500
might
look
like
in
that
scenario,
where
it's
really
the
AIS
fixing
resolving
all
the
way
through
the
process
even
deploying
the
code
through
its
devops
pipelines,
maybe
even
builds
the
devops
pipelines
to
deploy
to
cut
I,
don't
know
you
know
what
I
mean,
but
sort
of
you
know
really
that
that's
because
this
is
the
the
ultimate
the
Panacea.
Isn't
it
it's
the
Nirvana
of
AI
or
of
a
cognitive
business,
so.
E
Story
on
a
full
attack
and
where
it
touched
it's
it's
so
much
faster
and
then
and
then
the
humans
can
sit
and
down.
Look
at
it
and
figure
out
what
effects
it
had
in
their
Network
and
if
it
was
appropriate
or
not,
and
then
then
you
train
the
AI
to
change
its
process
based
on
how
you
want
it
to
to
manage
that
response.
Next
time,
interesting.
E
Sorry
and
just
to
add,
like
on
the
human
user
perspective,
the
easiest
way
for
humans
and
your
users
to
to
be
participated
is
from
like
an
Outlook
perspective
where
they
can
report
something
a
Spam
and
that
is
teaching
AI.
What
they
consider
spam,
what
they
don't
consider
spam
and
then
that
helps
your
artificial
intelligence
agent
in
your
network
as
well.
So
yeah.
D
I
keep
just
checking
that
yeah
I
was
just
I
know.
We
had
some
chat
about
the
kind
of
level
100
and
the
the
no
AI
there
and
I.
Don't
know
if
you
that
I
don't
think
anyone
has
no
AI,
even
if
they're
not
knowingly
using
it
on
there,
but
it
kind
of
linked
to
to
something
we've
been
talking
about
internally
in
terms
of
positioning
people
for
using
AI.
D
So
even
if
they're,
not
exploring
tools
and
making
of
it,
I
think
we
saw
with
things
like
delve
and
this
kind
of
goes
out
on
me
to
the
information
architecture,
which
probably
classifies
into
other
things,
but
positioning
your
content
in
a
way
to
be
able
to
use
AI,
making
sure
you're,
not
surfaces,
problems
that
you
have
in
there.
That
kind
of
fits
with
that
strategy
for
AI
and
your
almost
your
policies
around
it
and
the
reason
I'm
kind
of
mentioning
that
is
I
kind
of
feel
that
that
needs
to
come
a
bit
higher
up.
D
You
know
planning
what
you're
going
to
do
about.
It
should
be
further
focused
and
then
linking
with
those
implementations
as
the
further
maturity
of
actually
doing
things
with
it.
Yeah
and
I
also
feel
that,
with
the
500
we
have
it's
kind
of
not
leaving
so
much
growth.
For
what
interesting
new
things
come
out
that
we
haven't
thought
about
as
well.
B
Well,
wouldn't
wouldn't
each
sort
of
new
class
of
AI
each
each
new
capability
that
comes
out
sort
of
go
through
the
same
maturity
ladder
in
a
sense
I
mean
if
you
think
about
you
know
how
did
people
react
to
delve
in
the
beginning?
Well,
you
know
this
graph
stuff,
it
scared
everybody.
You
know
people
tried.
B
It
yeah
they
can
see
my
stuff,
and
so
you
know
the
maturity
level
of
what
we
see
as
the
graph.
Now,
as
you
know,
I
think
as
a
as
a
as
a
general
user
base,
we've
gone
up
that
ladder
and
then
something
else
comes
along
and
we
sort
of
start
at
the
bottom
again
and
then
you
know
so
yeah,
it's
almost
you
know,
I
think
I.
Think
an
organization
just
like
with
the
different
competencies
an
organization
might
be.
B
You
know
very
mature
with
a
certain
kind
of
AI
and
have
not
touched
something
else,
for
example,
if
if
you
think
it
it's
hard,
I
think
part
of
what
you're
you're
bringing
up
is
you
know?
Where
do
you
draw
the
line?
What's
Ai
and
what's
machine
learning
and
you
know,
was
Microsoft
graph,
AI
ish
sure
felt
like
it.
Didn't
it
so.
D
I
I
mean
you
know,
making
recommendation
based
on
things.
You
did
without
you
being
able
to
kind
of
directly
influence
that
and
also
think
the
the
within
this
we
don't
cover
so
much
the.
If
we
go
back
to
that
delve,
it's
the
just
turn
it
off
and
stop
it
happening.
I
mean
maybe
that
that
comes
back
to
what
the
no
AI
is,
and
maybe
it's
beyond
that.
It's
blocking
AI
avoidance.
D
C
D
C
Yeah
yeah
yeah,
yeah,
yeah,
yeah,
yeah
I
think
that's
right,
yeah,
yeah,
I
think
I
think
that's
right
and
and
yeah
200
that
way
or
or
we
start
to
play
around
and
try
some
of
these
things.
You
know
try
Cortana
or
try
insights,
yeah
via
insights.
You
know
that
kind
of
thing,
so
yeah
and
then
you've
been
yeah.
F
I
mean
that's
okay,
when
they
were
talking
about
about
the
sort
of
the
risks
and
the
people
side
of
things,
and
we
do
perhaps
want
to
take
a
bit
of
a
sort
of
techno
utopianism
type
approach
here.
But
the
fact
is
that
people
will
be
getting
hold
of
these
tools,
they'll
be
using
them.
Mistakes
will
be
made
and
they'll
be
compounded
and
there'll
be
all
kinds
of
issues
around
that,
and
it's
not
just
the
the
technical
literacy.
A
It's
a
really
good
point
and
something
we've
tried
to
touch
on
in
the
competency
is
talk
about.
You
know,
understanding
the
capabilities
and
constraints
of
the
AIS,
the
compliance
and
other
issues,
but
also
the
thing
about
having
sort
of
Audits
and
input,
output,
validation
and
a
whole
bunch
of
other
sort
of
feedback
routes
to
make
sure
the
AI
is
is
doing
what
you
intended
not
unintended
things
which
is
kind
of
exactly
the
same
as
you
have
with
people
again.
You
know
we
need
to
monitor
people,
so
we
should
be
monitoring
the
AI.
D
Okay,
I'll
try
and
dig
out
the
Tweet.
There
was
a
thing
yesterday,
one
of
the
first
things
when
gbt4
was
announced
was
they
have
a
set
of
tests
to
see
if
it
can
become
sentient
and
break
out
the
barriers
which
was
kind
of
fascinating
I'll,
see
if
I
can
dig
the
link
to
that,
because
it's
really
interesting
that.
E
A
A
C
Might
it
might
be
we
interviewed
them,
we
might
interview
them
and
say
you
know,
you
know:
we've
defined
these
types
of
behaviors,
here's
a
good
behavior,
here's,
a
okay,
Behavior,
here's,
a
bad
behavior
and
interview
them
and
see
which
one
they
come
up
with
and
if
they
say
the
bad
behavior.
Well,
that's
not
that
AI
is
not
right.
You
know
it
might
be.
That's
that's
how
it
sort
of
comes
about.
F
Things
too
much,
but
just
in
the
the
circumstances
that
Microsoft
actually
there
is
something
fundamentally
wrong,
with
the
way
that
their
AI
is
going
to
work,
that
it
starts,
leading
some
business
downturn
or
failures
and
those
kind
of
things
within
and
that
there
are
so
many
companies
with
their
bus
with
their
eggs
in
a
single
basket.
What
on
Earth
could
we
do?
What
could
what
kind
of
conversation
we
can
have
around
making
sure
that
people
are
able
to
make
sure
that
they
are
making
the
best
decisions
for
their
organization
yeah.
A
So
that
is
also
something
which
I
was
fretting
over
a
bit
over.
The
last
couple
of
days
is
where,
if
you've
got
organizations
that
have
allowed
the
AIS
to
be
the
dominant
voice
at
the
board
table,
then
you
know
how
do
they
want
to
stop
the
AI
started
to
game
against
each
other
and
and
the
whole
system?
Imploding
and
I
have
no
no
good
answer
to
that.
To
be
honest,
I
mean
we've
seen
some
of
this
stuff
potentially
happening
in
happening
in
global
Finance
markets.
A
F
C
There's
great
photo,
though,
for
the
this
is
all
great
photo
for
the
the
competency.
That's
you
know
this
is
you
know
things
which
really
should
be
level
300?
You
know
shouldn't
shouldn't
they
you
know
thinking
about,
maybe
maybe
not
the
protecting
from
you
know.
Global
demise,
but
just
to
have
you
know,
having
audits
having
those
checks
and
balances
yeah.
A
I
mean
yeah
at
the
risk
of
me,
going
off
on
one
of
my
pet
subjects.
I
think
you
know
one
of
the
things
that
will
emerge
from
this
as
we
try
and
nail
down
what
we
want.
Our,
how
our
AIS
are
meant
to
behave
is
it
will
make
us
thinking,
think
about
that
ourselves
inside
our
organizations
and
really
try
and
nail
down
what
we
think
is
and
isn't
acceptable
and
I.
A
You
know,
I
would
say
that
apple
and
Microsoft
will
have
different
values
and
therefore
their
AIS
will
behave
differently
and
we'll
then
see
it
again
at
a
national
level,
the
us,
but
how
you
will
behave
differently
to
you,
know
the
UK
or
to
China,
and
so
there's
all
these
things
that
you're
going
to
cut
coming
out
of
the
woodwork
that
AIS
will
be
a
tool
for
us
to
observe
as
a
as
well
as
being
a
tool
that
responds
to
those
different
differences
anyway.
Yes,
Aaron.
G
But
then
it's
in
Covenant
incumbent
in
the
user
on
how
to
manage
them
appropriately
and
if
we
are
in
the
lower
levels,
that's
a
very
difficult
thing
to
put
on
their
shoulders,
since
they
don't
have
that
framework
in
the
beginning
to
even
understand
how
these
are
really
working
and
so
I
love
hearing
that
you
guys
are
reaching
out
to
provide
guidance
and
saying
that
you
should
have
auditable
explainable
and
ethical
AI,
but
going
a
step
further
into
what
that
really
means
and
I
know.
You
mentioned
earlier,
it's
hard
to
codify
that
or
codify
that.
G
But
we
need
to
start
somewhere
with
that,
because
I
think
that
may
be
a
lot
of
people's
struggles.
Breaking
from
the
100
to
300
range
into
400
and
500
is
having
that
leg
up,
because
it
will
be
very
foreign
and
poorly
understood.
I.
Think
that
also
coupled
with
extreme
transparency
around
data
privacy
and
data
usage-
that's
especially
in
regulated
Industries,
that's
going
to
be
a
huge,
prohibitive
boundary.
G
This
is
kind
of
new
and
we're
letting
these
awesome
Technologies
out
there
without
a
lot
of
you
know
boundaries
other
than
just
well,
you
know
make
sure
you
guys
are
using
them
responsibly,
but
I
think
we
really
do
need
to
band
together
and
rapidly
produce
prototypes
of
of
that
codification
of
that
usage
and
start
iterating
on
that,
because
we're
going
to
be
experimenting
with
a
tool.
We
should
be
rapidly
experimenting
with
that
regulation
as
well.
Yeah
yeah.
A
We've
touched
on
that
perhaps
they
need
more
exploration,
but
we've
touched
on
the
need
to
do
that
within
the
model,
without
going
into
the
details
of
how
we
do
it
and
I
think
there's
definitely
some
work.
We
should
all
apply
ourselves
to
to
think
about.
How
can
we
set
those
standards
and
do
that
testing
and
so
I'm
going
to
be
making
myself
notes
about
some
of
this
stuff
now,
but
yeah
I
need
an.
Are
you
evil
test?
A
B
Well
and
they're
they're
things
that
are
evil
and
there
are
things
that
are
perceived
as
evil
too
so
you
know
does
do
the
people
in
an
organization
feel
comfortable
with
what's
happening
around
them
in
that
organization?
Is
on
that
Spectrum
right,
hey!
You
know
you
just
go
back
to
the
delve
thing:
hey
you're,
showing
my
stuff
I,
don't
like
that
that
felt
evil
to
a
lot
of
people.
G
A
B
A
I
think
we'll
skip
the
rest
of
this
slide,
I
guess
to
to
round
it
out.
We
have
our
usual,
you
know
set
of
links
and
thank
yous,
so
so
I'll
leave
that
up
on
screen
again
said
it
several
times.
I,
really
we're
going
to
get
this
model
published
very
rapidly,
and
then
we
really
would
like
people
to
go
and
poke
it
with
a
pointy
stick
and-
and
you
know,
tell
us
where
we've
got
things
wrong
or
we've
missed
stuff,
or
you
think
that
we
ought
to.
A
B
B
C
B
Probably
worth
mentioning
our
next
call
will
be
one
week
later
than
usual.
It's
highlighted
on
this
slide.
April
25th
instead
of
April
18th
I'm,
not
sure
if
David
got
the
the
invitation
updated,
but
I
think
he
has
so
just
make
sure
that
if
you're,
if
you've
got
it
on
your
calendar,
that
you
adjusted
if
it
hasn't
been.
A
Yeah,
good
and
otherwise
yeah
you've
got
multiple
ways
of
getting
in
touch
with
us.
Obviously,
you
can
directly
interact
with
documents,
but
you
know
tweet
either
to
us
or
about
us.
You
know
emails,
all
those
kind
of
things
all
available.
Email,
I,
remember
that
you
can
get
us
on
teams
actually
most
of
the
a
live
Federated
so
federatable.
So
you.