►
From YouTube: Latest on Project Cortex - Recap and updates
Description
This PnP community demo is taken from the monthly SharePoint Developer/Engineering update community call recorded on January 14, 2020.
In this 44-minute video, Naomi Moneypenny - Principal PM Manager, and members of her team Eric Gradel and Nkem Okoye, deliver an update on Project Cortex. First seen at Microsoft Ignite 2019, Project Cortex is the future of knowledge management for organizations and the result of a concerted effort by organizations across Microsoft to deliver a coherent cross platform experience for managing and organizing knowledge to empower people – just-in-time and in the context of their daily work. In Project Cortex, AI is fully integrated in this advanced ECM solution aiding with the evaluation of content as it comes into the organization to helping users find the right authoritative content when they need it to shape informed outbound communications.
Licensing is not yet defined for this premium offering that will be released in the first half of 2020. What is certain, is there will be plenty of extensibility opportunities for PnP Community members relating to Project Cortex.
Presenters:
• Naomi Moneypenny (Microsoft) | @nmoneypenny
• Eric Gradel (Microsoft)
• Nkem Okoye (Microsoft)
Supporting materials:
• Demos, videos, guidance at: https://aka.ms/ProjectCortex
• Content Services: https://aka.ms/sharepoint-contentservices
All SharePoint PnP Developer Community demos: https://www.youtube.com/playlist?list=PLR9nK3mnD-OWSbg0o9a7mx_E7s2u7h_o2
Learn more about the SharePoint developer community at: http://aka.ms/sppnp.
A
A
So
much
face
I
appreciate
it
actually
I'm
being
joined
by
two
of
my
colleagues
today
from
our
team,
so
Eric
Gradle
who's
in
our
knowledge
experiences
area
and
in
cambuquira
who's
working
looking
at
our
day,
zero
experiences
and
first-run
experiences,
so
Eric
is
actually
gonna
present
the
slides.
So
thank
you
for
doing
that
and
then
we'll
talk
through
a
couple
of
different
things
here.
So
I
know
we
have
a
limited
amount
of
time
and
I
know
that
both
of
ASA
and
I
can
talk
very
fast.
A
So
I
do
want
to
spend
a
little
time
here
to
go
ahead
and
introduce
project
cortex.
Really.
Hopefully
you
saw
with
some
of
the
announcements
that
we
had
at
ignite.
We
want
to
make
sure
that
you
feel
good
about
this
there's
different
areas
and
so
with
cortex.
Really,
you
know
we
think
about
sort
of
the
journey
that
we've
been
on
with
content
services
work,
obviously
a
long
history
and
sort
of
Microsoft
365
and
looking
at
the
content
services
area,
you
know
it's
really
about
understanding
how
we're
expanding
and
so
defining
that
vision
of
the
future.
A
So
what
does
the
content
services
the
traditional?
What
we
used
to
call
ECM
way
back
sort
of
thinking
about
how
that
evolves?
And
then
really
thinking
about
you
know,
what's
the
future
of
knowledge
inside
of
an
organization?
So
if
we
think
about
sort
of
where
we
want
to
go
with
with
project
cortex,
it's
really
around
these
three
big
areas.
So
looking
at
sort
of
how
we
manage
and
organize
and
empower
this
content,
so
when
they
think
about
management,
it's
really
about
you
know.
A
How
are
we
helping
you
to
connect
all
of
the
data
you
have
across
your
Microsoft
365
ecosystem
and
beyond?
We
do
of
course,
recognize
you
know.
Companies
are
connected
into
lots
and
lots
of
different
data
sources,
you've
great
deal
of
applications.
We
want
to
think
about
how
we're
able
to
take
a
lot
of
that
information
and
move
that
into
a
knowledge
network
and
be
able
to
use
all
of
that
data.
That's
available
to
you
in
your
ecosystem.
A
It's
really
about
you
know
how
we're
helping
you
to
really
get
that
information
that
you
need
that
knowledge
at
the
time
that
you
need
it.
So
how
do
we
make
sure
it's
embedded
in
all
the
experiences
that
you
have
in
Microsoft
365
and
that
we're
really
delivering
it
to
you,
where
you
need
it
every
day?
So
we
think
about
sort
of
the
next
stages
here
and
thinking
about
where
we
are
going
with
with
project
core-tex.
We
really
want
to
think
about
how
we're
putting
all
these
pieces
together.
A
So
you
know
leveraging
that
investment
that
we
have
of
a
see
connecting
up
what
we're
doing
with
sharepoint
and
then
understanding
a
little
bit
more
about
how
we
can
you
know,
make
the
best
of
AI
and
all
the
pieces
that
we
have
across
Microsoft.
This
is
not
just
about
us,
as
the
you
know,
onedrive
and
SharePoint
team.
This
is
really
about
how
we
deliver
across
the
full
breadth
of
Microsoft
and
making
sure
that
we
have
the
great
machine
learning
and
the
great
AI
available
and
make
putting
these
things
together.
And
then
you
know
expecting
this.
A
The
way,
oh
of
course,
part
of
the
trust
and
security
and
controls
that
we
have
with
Microsoft
365.
So
all
of
those
pieces
coming
together.
So
it's
kind
of
thinking
about
these
different
areas,
and
it's
all
good.
So
with
this
sort
of
empower
area,
it's
really
about
how
we
delve
into
the
knowledge
experiences,
and
so
it's
easier,
perhaps
to
start
from
thinking
about
what
an
end-user
sees
and
making
sure
that
you
know
they're
getting
the
information
that
they
need.
A
C
This
is
Eric
speaking
about
the
knowledge
management
area.
Knowledge
management
is
all
about
enabling
people
and
organizations
to
find
discover
and
connect
to
the
people
and
content,
that's
important
to
them,
and
so,
in
this
context,
what
you're,
seeing
is
what
we're
calling
annotations
highlights
so
core
has
been
highlighted.
C
C
We
can
see
an
example
of
a
topic
card
here.
A
quick
summary,
that's
AI
generated
is
appears
on
the
type
of
card,
along
with
the
related
people
and
resources,
and
so
this
is
a
good
way
for
someone
who
doesn't
have
context
about
core
to
to
get
the
the
quick
download
of
what
this
project
is
all
about,
and
the
reason
that
I
can
see
this
card
is
because,
as
the
user
I
already
have
permissions
to
read
the
files
and
pages
that
were
used
as
a
source
for
the
AI
mining.
C
So
here
we're
we
can
see
the
related
resources
and
people
again.
This
is
mine
from
pages
and
and
documents
that
were
used
as
a
source
for
this.
This
topic
card
creation,
and
then
this
demo
shows
what
this
looks
like
in
the
context
of
a
SharePoint
page
again,
project
core
is
highlighted
here,
along
with
project
soar.
These
are
both
topics
that
the
the
AIA
has
identified.
As
as
projects
that
may
be
relevant
for
your
work
and
anywhere
you
work
across
all
Microsoft
canvases.
C
And
then,
if
the
user
clicks
through,
they
will
get
a
full
topic
page,
and
you
can
think
of
this
as
a
sort
of
an
AI
generated,
wiki
page
which
which
has
more
description
on
the
topic
and
allows
that
allows
me
to
discover
related
topics
and
click
through
the
knowledge
network.
That's
been
created
via
the
connections
between
these
topics.
C
This
same
sort
of
information
is
also
available
as
part
of
Microsoft's
search.
So
if
I
search
for
project
core
on
on
my
company,
intranet
I
will
locate
the
same
page
and
again
just
more
description
of
you
know.
The
the
related
people
and
resources
appear
on
this
on
this
wiki
page,
and
this
shows
the
the
source
of
the
source
that
was
used
to
to
create
this
AI
generated
wiki.
C
These
are
all
the
the
documents
that
that
the
AI
has
has
found
our
related
and
mined
for
information
to
create
this
topic
page,
and
then
this
is
just
visualizing,
the
connections
amongst
topics.
So,
as
we
pointed
out,
the
AI
has
identified
project
core
and
projects
or
as
as
being
likely
project
names
and
has
created
a
relationship
between
these
names.
This
visualization
will
be
used
to
allow
people
to
to
navigate
amongst
related
topics
and
is
used
as
a
way
to
sort
of
serendipitously
discover
topics
that
you
know
are
in
in
the
sphere
of
project
core.
C
And
soon
we
expect
to
be
able
to
include
questions
and
answers
via
Yammer
on
this
wiki
as
well,
and
so
Yammer
will
will
be
a
source
that
we
ingest
information
from
to
augment
these
pages.
But
you
can
begin
to
see
how,
through
a
combination
of
webparts
that
appear
on
on
this
topic
page,
you
can
begin
to
create
a
rich
source
of
of
content
around
project
core.
C
Mark
8
is
an
example
of
another
project
that
the
AI
has
located,
and
this
shows
of
you,
the
users
individual
view
of
the
topics
that
are
in
their
sphere.
So
you,
the
user,
are
related
to
mark
8
in
projects
or
so
they
will
appear
on
what
we're
calling
the
knowledge
center
of
you,
and
this
is
a
new
destination
that
users
can
can
go
to
to
discover
all
the
topics
that
are
related
to
them
via
the
documents
that
they've
worked
on
and
also
discover
new
topics
that
might
be
of
interest
to
them.
C
You'll
also
use
this
view
to
to
manage
the
the
topics
that
you're
connected
to
and
make
simple
edits
as
well.
So
you
don't
need
to
go
to
this
destination.
To
discover
this
knowledge
than
the
knowledge
as
you've
seen,
will
be
populated
throughout
the
Microsoft
canvases
via
these
annotations,
and
so
you
can
serendipitously
happen
upon
this
knowledge.
C
But
if
you
are
curious
enough
to
go
and
see
what
are
the
other
topics
that
may
be
of
interest
to
me,
what
are
the
things
that
I'm
connected
to
this
would
be
the
place
to
to
go,
and
so
this
shows
you
know
all
the
news
posts,
all
the
other
topic
pages
that
may
be
of
interest
to
you.
That
way,
you
get
you
know
a
sense
of
what's
fresh,
what's
trending
that
may
be
related
to
what
you're
working
on.
C
And
as
I
pointed
out
a
little
earlier,
these
are
standard
SharePoint
pages,
meaning
that
you
can
customize
your
view
of
these
individual
pages.
Web
parts
can
be
added
as
needed,
and
these
knowledge
web
parts
that
we
are
demoing
can
also
be
added
to
any
other
standard,
SharePoint
page.
So
you
know
very,
very
flexible
and
in
how
you
present
and
consume
this
information.
We
we
expect
that
people
will
populate
these.
These
knowledge
web
parts
throughout
SharePoint
pages
as
they're
relevant,
and
this
shows
you
know
the
exact
same
topic
highlights
viewer
just
in
the
context
of
words.
C
So
you
know
before
we
were
an
outlook
and
we
saw
project
core
highlighted
here,
we're
seeing
it
highlighted
in
the
context
of
a
word
document
and
using
smart
lookup.
You
get
this
a
very
similar
view
of
the
the
topic
page
with
the
definition
and
related
people
and
resources,
and
we
expect
to
see
this
in
in
teams
as
well.
A
So
much
so,
let's
talk
about
the
next
couple
of
sections
here.
Just
keep
going
through
the
slides
era
can
help
we'll
keep
pushing
on
here.
So
let's
talk
a
little
bit
about
this
organizational
area
because
we
want
to
think
about.
You
know
the
content,
that's
coming
into
your
company
and
how
do
we
think
about
making
that
smarter
as
it
comes
in
so
as
Eric
mentioned,
these
great
rich
experiences
that
we
have
from
a
knowledge
perspective?
A
We
were
trying
to
really
make
that
simple
for
the
end-user
to
discover
and
to
be
able
to
get
to
the
information
that
they
need
being
able
to
understand
sort
of
what's
available
as
part
of
the
topic
and
everything
else.
But
we
also
want
to
think
about
how
a
business
process
can
be
smarter.
How
do
I
make
the
content?
That's
coming
into
my
company
a
little
bit
smarter,
and
so
there
are
lots
of
ways
to
do
that.
A
There
we're
also
looking
at
the
sort
of
extra
piece
here.
That
is
understanding.
What
is
the
content
inside
of
the
document
too?
So
you
look
at
the
connections
here.
We
look
at
the
signal
depth
and
then
we're
understanding
the
content
itself
and
that's
how
we're
actually
extracting
a
lot
of
that
topical
information.
So,
if
you
think
about
it,
we're
using
a
combination
of
different
methods,
so
they
think
about
the
Microsoft
graph.
We
have
all
that
great
layer
of
data
and
all
the
content.
That's
it
that's
being
stored
there.
A
You're
collaborating
the
documents
or
meetings
or
tasks
all
those
pieces,
and
then
you
can
think
about
the
next
layer.
On
top
of
that
is
really
understanding
a
little
bit
more
about
how
we
can
make
connections
between
those
things,
because
we
can
get
more
insight
by
processing
it
in
different
ways.
Now
what
we're
doing
is
actually
looking
through
a
lot
of
the
knowledge
that
we
have
inside
of
organizations.
A
So
the
interesting
thing
here
is
we're
actually
able
to
use
probabilistic
programming
to
understand
some
of
the
different
areas
of
documents,
understand
the
content
itself
and
the
richness
of
the
content
and
then
look
for
ways
to
sort
of
verify
that
those
connections.
When
we
do
that
verification,
then
we
can
actually
do
analysis
across
those
signals
and
then
be
able
to
take
that
information
to
understand
how
things
are
connected
and
new
in
different
ways.
A
And
so
it's
really
interesting
to
take
the
approach
that
says
how
can
I
see
if
a
project
and
topics
are
connected
together
without
somebody
actually
formally
telling
me
that
well,
I
can
look
through
this
kind
of
network
of
connections,
the
network
of
signals
that
we
have
and
explore
new
edges
to
understand
how
things
are
related
together,
and
so
that's
really
understanding
how
we
can
create
these
topical
Maps.
So
we
just
showed
but
also
thing
about
more.
The
richness
of
the
signals
that
we'll
have
in
the
future
too
so
I
think
I'm
the
organized
side.
A
Here
we
have
a
little
bit
more
in
terms
of
understanding
the
controls
that
are
part
of
the
system,
and
so
with
the
content
controls
themselves,
we
had
lots
of
questions
at
ignites
and
beyond
so
I
think
it's
it's
worth
to
sort
of
talk
through
some
of
these
areas.
So
people
are
concerned,
obviously,
with
some
of
the
past
experiences
they
may
have
had
use
sort
of.
How
do
we
make
sure
that
we
feel
good
about
the
ways
that
we're,
including
this
information,
so
first
off,
obviously
with
Microsoft
365?
A
We've
got
a
couple
of
different
things:
all
the
IP
and
the
compliance
and
security
controls
that
you
expect
with
Microsoft
365
and
then
we're
adding
a
layer
on
top
of
that,
especially
for
this
knowledge
mining
area.
So
how
do
I
think
about
my
knowledge
network?
It's
really
about
how
I'm
able
to
allow
different
sites.
So,
for
example,
when
you
first
start
with
project
core-tex
you'll
probably
want
to
scope
it
through
a
certain
number
of
sites
that
you
feel
good
about.
A
You
can
understand
and
have
that
information
traversed
and
mindful
so
we
have
obviously
allow
sites
we'll
have
things
like
block
sites
in
there
you'll
be
able
to
exclude
content
based
on
sensitivity
labels
so
again
that
great
connection
with
the
Microsoft
IP
work,
that's
been
going
on
and
then,
of
course
we
are
observing
the
search
setting.
So
no
Index
are
no
crawls.
A
So
if
you've
ever
had
experiences
before
where
you've
had
content
that
maybe
showed
up
in
delve
in
the
past-
and
you
didn't
want
that
to
happen-
we
will
obey
those
same
Microsoft
search
settings
as
well
and
then
with
topics
themselves.
So
we
actually
be
able
to
give
you
a
way
to
visualize
this,
as
an
administrator,
so
being
able
to
block
specific
topics.
You'll
be
able
to
see
the
different
topics
that
come
in
so
think
about.
A
The
connections
that
we
have
we've
managed
metadata
you'll
also
be
able
to
use
things
like
a
list
of
terms
or
different
acronyms,
and
then
we
can
also
manage.
Who
sees
those
topics
too.
So
kind
of
an
interesting
piece
here
is
thinking
about
how
you
do
adoption
inside
of
the
organisation's.
So
at
first
you
know
maybe
want
a
small
group
of
users
using
this
and
then
think
about.
A
You
know
the
different
ways
that
we
can
have
all
of
the
things
that
could
possibly
be
a
candidate
topics,
so
think
about
all
of
the
work
that
it's
doing,
cortex
is
going
through
its
mining.
All
these
project
names
topic
names
and
then
it
can
actually
show
you
a
list
of
all
of
those
candidates,
and
so
we
may
want
to
do
some
subject
matter:
experts
to
verify
those
candidates
first
or
whether
you
choose
to
have
your
end
user.
A
A
population
see
all
of
the
things
that
have
already
been
confirmed
as
topics
and
so
there's
a
kind
of
different
and
adoption
and
ableman
options
that
we
have
there
and
the
same
thing
in
this
August
topic
manager
area,
because
you
think
about
you
know
we're
going
to
show
you
some
suggestions
of
project
names
and
topic
names
that
we
found
and
then,
when
you
say
hey,
is
this
a
real
topic
red?
So
we've
done
a
lot
of
work
in
terms
of
feedback,
but
we
also
want
to
take
a
step
before
that
and
say
hey.
A
You
know
what
what
does
me
as
an
administrator
me
as
a
knowledge
manager
like?
Does
this
really
look
like
a
good
candidate
topic?
Is
it
something
I
can
view
and
confirm
and
then
once
I've
done,
that
I
have
the
ability
to
have
that
be
available
to
my
end-users,
and
so
the
the
goal
around
this
is
to
really
build
up
the
adoption
and
the
enablement
of
what
knowledge
management
look
like
inside
of
your
organization.
A
What
does
it
look
like
from
the
subject
matter,
expertise
where
area
and
then
how
do
we
make
sure
that
people
are
having
a
great
experience
from
day
one?
So
if
you
take
a
little
bit
of
time
to
do
some
of
the
work
to
say?
Yes,
these
are
real
topics.
These
are
real
project
names.
These
are
things
that
I
want
to
make
sure
my
end-users
are
seeing
then
it'll
be
a
great
experience
for
all
of
them,
and
then
these
abilities
around
sort
of
merging
and
managing
topics
so
that
whole
topic
lifecycle.
A
We
realize,
of
course,
project
names
and
topic
names
change
inside
of
our
organization.
We
want
to
be
able
to
manage
those
things
and
then
the
ability
to
curate
those
topics
as
well
so
making
sure
again
for
all
the
different
organization
styles,
and
we
realize
that
even
within
different
business
units,
there
can
be
a
different
style.
So
some
folks
may
have
a
more
restrictive,
save
style
and
then
other
areas
of
the
business
may
be
more
open,
so
we're
trying
to
have
the
range
of
controls
that
allow
you
to
implement
cortex
the
way
that
you
want
it.
A
So
with
that,
let's
move
on
to
the
move,
the
connection
area
here-
let's
just
keep
going
through
and
so
we'll
talk
through
now
a
little
bit
more
about
content
understanding
and
how
we
can
drive
intelligence
in
there.
So
we
think
about
the
knowledge
network
is
that
information
we're
mining
through
always
inside
of
your
organization,
we're
understanding
that
contents.
You
know
we're
figuring
out
sort
of
what's
inside
of
the
documents,
and
so
when
I
mentioned
earlier,
it's
a
little
bit
more
about
the
business
process,
information
and
how
it's
coming
in.
A
So
we
want
to
make
content,
that's
actually
coming
into
your
organization's
smart
or
to,
and
we
want
to
give
you
great
ways
to
be
able
to
manage
that
and
extract
the
metadata
that
you
need
from
it
too.
So,
let's
go
through
and
I
think
we're
going
to
go
through
different.
The
different
methods
here,
so
they'll
have
three
different
methods
as
part
of
the
project
or
tax.
A
A
Processing,
and
this
is
about
conjunction
with
our
partnership
with
the
Microsoft
power
platform,
which
is
a
new
tool
called
a
eye
builder,
and
so
you'll
be
able
to
build
unsupervised
and
supervised
machine
learning
models
for
processing
forms,
and
this
is
really
about.
Having
information
looks
very
similar
to
each
other,
so
think
of
it
as
a
template.
You're
filling
out
forms
you've
got
that
kind
of
information.
It's
easily
extracted
from
a
standardized
template,
that's
a
great
way
to
be
able
to
process
that
information
and
then
for
unstructured
documents
like
contracts
and
those
business
agreements
or
proposals.
A
So
let's
talk
through
some
of
the
the
areas
here
so
I
know
we
have
a
quick
demo
here
in
terms
of
the
automated
object
detection,
so
the
one
of
the
things
we're
doing
is
creating
a
new
destination
for
content,
sight,
centers
and
content
centers
initially
for
the
first
version
of
project.
Cortex
will
be
the
place
where
you
can
do
machine
teaching,
but
we'll
also
have
all
of
these
other
analytics
capabilities.
A
We'll
also
have
a
solution
solution
accelerator
available
too
for
analytics,
so
you'll
be
able
to
understand
a
little
bit
more
about
the
informations
being
processed
through
your
pipeline,
so
they'll
be
the
new
since
center
and
if
we
click
through
here
you'll
see
that
as
I
upload
different
documents,
you'll
see
that
I
can
put
in
things
like
engineering,
diagrams
I
could
upload
PDFs,
etc,
and
so
we
want
to
make
sure
that
it's
actually
smart,
out-of-the-box,
and
so
in
this
case
you
can
see
I've
got
a
PDF
documents.
I've
got
an
engineering
wireframe
here
and
oftentimes.
A
You
know
what
those
kind
of
drawings
we
really
want
to
just
extract
the
handwriting
on
there
as
well.
So
imagine,
for
example,
is
a
wireframe
diagram
or
it's
a
return
or
something
and
a
customer's
marked
in
handwriting
or
something
failed.
Then
I
can
upload
that
and
as
I
upload
and
click
through
here
then
you'll
see
that
we
are
able
to
extract
that
metadata
automatically
from
there.
So
I
upload
those
things
normal
way.
A
We
do
things
in
SharePoint
and
then
be
able
to
click
through
and
then
make
sure
that
the
handwriting
is
being
extracted
from
those
documents
and
so
that
populates
the
column
columns
here
inside
of
my
document
library,
it's
kind
of
an
example
of
terms
of
doing
that
kind
of
smart
against
smart
out-of-the-box.
We
want
to
make
sure
that
things
like
handwriting
extraction
are
built
in
being
able
to
extract
the
pieces
that
are
most
interesting,
so
I'm
gonna
hand
it
over
to
you
in
cambuquira.
D
What
are
some
of
the
key
value
pairs
and
extract
them,
while
this
is
really
cool
is
because,
instead
of
you
to
actually
manually,
go
through
each
document
to
extract
less
you
like
the
vendor
name,
the
shipping
information,
once
you
upload
the
documents
here
on
this
new
sort
of
content,
side
collection
that
we
have
content,
you
can
definitely
you
can
have
the
model
applied
on
top
of
the
documents
to
extract
metadata
from
it.
So
here
we've
integrated
ai
builder,
with
SharePoint,
so
here's
an
example
of
you
know
purchase
orders.
D
So
here
we
were
expecting
you
to
upload
five
documents
with
the
same
layout
to
sort
of
train
the
model.
So
let's
say
you
uploaded
a
purchase
order.
Here
is
an
example
of
the
document
you
uploaded
and
then
you
can
see
the
model
sort
of
goes
through
to
identify
some
key
value
pairs,
the
names,
the
purchase
other
ship
to
sender,
and
then
you
can
see
how
some
of
those
selected
fields
will
then
show
up,
and
then
you
have
the
ability
ability
to
either
change
what
you
want
the
field
name
to
be
or
really
confirm
and
say.
D
D
The
model
then
goes
through
each
of
the
documents
and
then
extracts
automatically
some
of
that
information
that
you've
sort
of
identified
that
you
need
out
of
the
documents
so
I
think
it's
pretty
cool,
because
these
are
things
that
we
know.
People
did
manually
back
in
the
days
and
it
really
helps
you
in
case.
You
need
this
sort
of
information
to
see
through
other
processes
or
then
just
apply
specific
rules
on
top
of
it.
D
Another
thing
that
we're
introducing
as
part
of
project
cortex
is
really
ways
to
identify
file,
file
types
and
then
extract
entities
from
those
types
of
files.
So,
while
the
AI
do,
that
is
really
good
for
structured
information
and
similar
machine
teaching.
The
machine
teaching
capability
that
we're
introducing
helps
subject
matter
experts
to
really
look
at
documents
that
could
be
in
different
formats.
D
So
this
is
a
little
bit
it's
more
involved
than
the
AI
building
stuff,
because
the
AI
building
has
you
know
those
definitive
templates,
but
this
is
really
cool
because
you
can
extend
it
to
different
types
of
document.
Sizing
you're,
not
limited
by
you
know.
Are
they
all
PDF
sources?
Are
they
all
Word
documents?
So
here,
as
I
mentioned,
we're
enabling
subject
matter
experts
to
teach
the
machine
which
is
pretty
cool
because
then
you
have
control
over
what
is
or
what
isn't.
D
So
you
first
of
all
upload
some
positive
and
negative
examples,
and
then
we
expect
you
to
really
help
the
model
say
like
hey
yeah.
This
is
what
it
is.
Maybe
add
some
annotation,
it's
you.
If
you
want
and
then
what
we
also
help
to
do
is
to
also
publish
a
confidence
score.
So,
as
you
continue
uploading
certain
types
of
positive
documents
and
negative
documents,
you
can
really.
You
can
really
monitor
how
well
the
model
youth
train
is
doing
with
extracting
that
information
from
the
documents.
D
Once
you
have
identified
model
you
can
then
publish
it
to
a
SharePoint
library
and
and
also
integrate
some
of
the
metadata
that
gets
extracted
into
some
of
the
regular
flows
that
you
do
today.
So
you
can
kick
off
different
workflows
for
people
based
off
of
what's
been
extracted
and
from
the
documents.
A
Great
thanks
so
much
thank
him,
so
it
was
great
to
hear
these
different
areas
and
sort
of
think
about.
You
know
the
difference
between
what
we
would
be
able
to
do
with
forms
processing,
as
you
mentioned
with
that
structured
data,
and
then
the
machine
teaching
approach
which
is
really
unstructured
data,
and
then
it
was
we're
extracting
all
of
that
metadata.
A
Hopefully,
you
realize
that
metadata
has
become
even
more
important
in
all
of
these
different
services
and
so
that
extracted
metadata,
of
course,
lives
in
columns
and
content
types
and
being
able
to
use
that
across
the
different
areas.
So
if
we
go
to
the
next
slide,
we
can
take
out
some
of
the
different
ways
that
we've
been
improving,
metadata
and
thinking
about
all
these
different
areas.
So
you
look
at
the
ways
that
were
ingesting
content
inside
of
your
organization.
We
want
to
think
about
how
we're
enriching
that
knowledge
network
as
it
comes
in
now.
A
A
So
we're
making
a
whole
bunch
of
different
investments
here,
because
we
want
to
really
want
to
transition
the
information
that's
coming
in
as
part
of
the
information
business
process
into
how
we
can
transform
it
into
knowledge
inside
of
your
organization
and
then
how
we're
really
helping
to
convert
that
into
learning
as
well.
So
it
really
starts
off
with
some
of
these
smart
ingestion
capabilities
so,
as
in
CHEM
showed
you
thinking
about
forms,
processing
and
you
about
machine
teaching
where
you've
got
subject
matter
experts
who
are
able
to
teach
the
way
that
they
understand.
A
You
know
whether
it's
legal
clauses
inside
of
a
document
or
our
project
management
or
an
RFP
pieces
is
being
structured,
and
then
we
think
about
you
know
using
these
different
ways
of
machine
learning
and
machine
teaching
to
help
with
that
classification
area.
And
then
we
really
want
to
be
able
to
deliver
more
insight
as
part
of
that.
So,
if
you
think
about
the
transition
between
what
becomes
knowledge
more
becomes
learning
inside
of
a
system,
it's
really
about
how
we're
able
to
do
a
next
level
of
insights
and
analytics
on
top
of
it.
A
So
bringing
all
these
pieces
together.
So
if
you
next
slide
and
we'll
go
into
sort
of
some
of
these
new
experiences
that
we
have
for
metadata
so
of
many
folks
who
have
been
in
the
community
for
a
long
time,
we'll
remember
the
managed
metadata
service
from
SharePoint
2010
onwards,
and
so
we're
really
taking
that
service
investing
heavily
in
it
and
making
it
something
that
becomes
suite
wide,
so
not
just
SharePoint
anymore,
but
really
taking
it
into
Microsoft
365,
and
so
we've
had
customers.
A
A
A
We
also
think
this
is
really
important
for
other
applications
may
be
things
like
sa
Peaks
after
where
you
already
may
be
doing
some
master
data
management
another
place,
and
you
want
to
keep
some
of
those
pieces
synchronized
as
well,
and
then
we
think
about
content
types.
So,
as
I
mentioned
earlier,
we
are
investing
heavily
in
content
types
as
well.
A
And
the
same
thing
is
true
on
the
content-type
piece
here,
because
we
want
to
be
able
to
make
sure
that
you're
able
to
look
at
the
parent
content
types
sort
of
be
able
to
put
those
things
together
and
then
we
want
to
think
about
the
ways
that
were
categorizing
these
as
well,
and
then
you
can
think
about
how
those
get
pushed
to
a
hub
site
themselves.
We're
also
adding
end-user
controls,
because
we
want
to
add
content
types
to
document
libraries.
A
So
an
end
user
can
say,
add
a
content
type
because
it
makes
sense
just
to
do
it.
There
sort
of
think
about
that
template,
past
policy
and
having
that
will
be
part
of
the
document
library
as
well,
and
then
thinking
about
you
know
these
end
user
experiences
here
so,
for
example,
in
this
in
this
area.
We're
really
thinking
about
you
know
how
we
can
add
those
content
types,
the
different
columns
here
and
again
behind
the
scenes,
things
like
machine
teaching,
they're
actually
using
content
types
to
store
that
data.
A
So
you
think
about
you,
know:
I've
trained
a
business
process.
Maybe
I
have
something
like
you
know,
proposals
that
are
coming
into
my
organization,
I'm,
extracting
the
metadata
there
that's
actually
getting
stored.
Is
you
know,
part
of
that
content
type
behind
the
scenes
and
so
being
else
attach
all
the
different
things
that
I
would
be
able
to
do
with
content
types
so
about
policy
information.
Think
about
the
way
it's
being
syndicated.
Think
about
you,
know
the
workflow
and
other
things
that
I
want
to
have
as
part
of
that
too.
A
A
You
know
I
should
be
able
to
easily
filter
the
documents
within
those
libraries
to
be
able
to
see
those
things
by
category
in
this
case,
so
seeing
all
of
those
different
areas
from
the
term
set
perspective,
and
then
the
last
area
here
just
sort
of
make
sure
we
understand
a
little
bit
more
in
sort
of
just
thinking
about
you
know
the
the
foundational
investments
that
we're
making
here.
So
we're
really
thinking
about
this
is
very
coherent
experience
across
Microsoft
365.
A
So
again,
the
the
first
version
of
project
cortex
are
making
these
these
big
set
of
investments,
but
you'll
see
even
more
of
those
things
coming
across
Microsoft
365
and
the
suite
wide
experiences.
We
also
want
to
think
about
how
you're
taking
these
enterprise
content
types
and
how
you
think
about
that
based
on
terms
so
there's
the
search
functionality.
A
That's
thinking
about
you
know
the
ways
that
we're
doing
knowledge
finding,
but
we
also
want
to
have
a
managed
set
of
metadata
as
well
for
organizations
and
so
finding
that
balance
between
there
and
making
sure
that
we
can
help
users
to
be
able
to
discover
the
content
that
they
need
based
on
some
of
those
areas
and
they're.
Thinking
about
you
know
that
consistent
tagging
experience,
because
we
want
to
think
about
how
you're
doing
term
suggestions
and
auto
tagging
and
those
all
of
those
different
areas
coming
together.
A
So
you
know
you
saw
earlier
in
the
knowledge
Network
demo
that
Eric
did
sort
of
seeing
that
area,
for
example
where
teens.
As
being
you
know,
picking
up
on
the
topic,
that's
being
mentioned,
you
know
being
able
to
see
that
automatically
being
able
to
do
things
like
smart
lookup
inside
of
word
those
kinds
of
areas
and
then
thinking
about
how
we're
doing
that
water
tagging
essentially
for
you
as
well
and
then
the
last
area
here
around
enterprise
content
types.
A
You
know
we
want
to
think
about
the
metadata
that's
coming
across
your
entire
tenant
and
to
be
able
to
manage
that
in
a
much
more
cohesive
way
as
well.
Okay,
so
I
think
just
the
last
couple
of
slides
here,
sort
of
you
know
just
recapping
a
little
bit
here,
thinking
about
the
different
ways
that
we're
managing
all
of
this
information
inside
a
project
cortex
applying
some
really
amazing,
AI
capabilities.
If
I
think
about
you
know
all
of
that
content.
A
That's
coming
into
your
organization,
we're
really
taking
a
new
set
of
capabilities,
probabilistic
area,
for
example,
to
do
some
of
their
knowledge
mining,
then
we're
layering
on
new
capabilities
around
deep
neural
networks
and
some
deep
learning
there
we're
understanding.
You
know
the
different
ways
that
projects
and
topics
and
that
language
inside
of
an
organization
and
how
that's
being
things
are
being
talked
about,
so
that
we
use
that
as
infrared
signals
and
then
being
able
to
obviously
couple
that
with
all
of
the
great
capabilities
that
we
have
across
Microsoft
365
and
then
really
thinking
about.
A
You
know
how
we're
able
to
organize
and
then
of
course,
empower
folks
here
so
using
those
great
content,
intelligence
content,
understanding
pieces,
making
sure
we
can
actually
get
intelligence
from
inside
of
the
content
as
it
comes
into
your
organization
and
then
really
making
that
part
of
that
knowledge
network
going
forward
so
that
we
can
deliver
it
inside
of
great
experiences.
And
then
just
you
know
a
reminder
where
to
go
to
get
more
information
about
project
cortex.
A
We
are
scheduled
for
general
availability
in
the
first
half
of
this
year,
so
project
team
is
working
very,
very
hard
on
all
of
these
pieces
and
then
we
do
keep
their
Resource
Center
up
to
date
as
well.
So
if
you
go
to
this
link,
you'll
actually
see
that
there
is
a
you
scroll
down
the
page,
there's
a
few
different
forms
and
things
you
can
fill
out.
A
B
B
A
B
And,
and
also
just
to
recap,
the
timeline-
the
h1
2020
means
spring
this
year,
so
within
the
upcoming
months,
but
because
there
was
a
lot
of
questions
related
on
when
it's
X&Y
on
see
available.
When
is
this
available
when
it
is
available,
no
timelines,
yet
we
can't
give
you
an
ETA,
and
this
is
a
kind
of
a
classic
discussion.
B
Everybody
keeps
on
always
asking
about
the
ETA,
but
we
don't
want
to
promise
to
deliver
anything
by
end
of
as
an
example
by
end
of
this
month,
because
it
what
if
we
should
sleep
and
then
you
make
your
commitments
as
a
customer
on
that
promise,
and
then
you
will
be
mad
and
that's
why
we
don't
give
a
date
so
until
we
can
guarantee
that
today
and
be
kept
answer,
the
question
can
we
say,
share
point
with
project.
Coders
concept
can
offer
advance
CMS.
That's
that's
a
good
question.
Actually,
so
is
this
project
code
access?
B
A
I
think
it's
a
great
thing.
It's
really
was
one
of
the
genesis
areas
for
project
cortex.
We
looked
you
know
at
the
content
services
area,
we
spent
a
long
time.
You
know.
Looked
at
the
Gartner
Magic
Quadrant.
We
looked
at
always
again
of
different
areas
and
industry
reports
and
we
really
try
to
look
at
like
what
could
we
do
this
that's
new
and
different,
and
so
we
look
to
those
areas.
We
spend
a
lot
of
time,
obviously
talking
to
customers
and
doing
doing
research
there,
and
so
the
advanced
CMS
capabilities.
A
You
know
it's
so
daunting
to
organizations
to
say
I'm
gonna
go
through
and
find
you
know,
here's
all
the
projects
in
my
company,
here's
all
the
topics
that
could
possibly
be
in
my
company
right,
it's!
It
is
its
enormous
effort,
and
so
you
know
our
conclusion
was
you
know
to
really
and
offer
these
advanced
services
that
we
should
be
applying
AI?
What
they
are
is
good
for.
You
know
importa
worked
in
ml
for
the
last.
You
know
15
years
at
this
point,
and
so
you
know
what's:
what's
it
really
great
at
it's
really
great.
A
You
know
brute
force
coming
through
a
whole
bunch
of
things
and
giving
you
a
great
first
cut,
and
so
we
want
to
make
sure
that
we're
marrying
both
of
those
things
together
that
the
advanced
capabilities
we're
using
these
new
machine
learning
techniques
we're
using
some
really
cool
tech
from
across
Microsoft.
But
we
also
want
to
make
sure
there's
lots
of
space
in
there
for
when
people
need
to
curate
and
add
their
subject
matter
expertise
in
there
and
to
make
it
easy
to
keep
those
things
up
to
date.
So
that
really
was
the
genesis.
This
whole.
B
A
B
And
it
is
absolutely
on
and
it's
kind
of
a
recapping
on
that
for
now.
This
is
the
time
when
we
truly
truly
truly
take
advantage
of
the
technique
at
the
the
awesomeness
of
rushar
and
with
the
technologists,
so
I
sure
has
a
lot
of
these
technical
bits
and
pieces,
but
there
are
individual
tools
and
individual
things
or
it's
a
nail
or
it's
a
it's
a
hammer,
and
but
we
now
as
part
of
Microsoft
resist
if
I
we
combine
and
use
that
and
provided
that
technology
and
provide
that's
easily
available
for
everybody
to
consume.
B
Now
there
was
two
other
things
related
on
what
I
just
said
and
what
you
know
me
said
related
on
taking
advantage
of
truly
the
cloud
and
the
flexibility
of
the
cloud.
Obviously,
then
that
has
an
implications
for
on-premises,
because
we
do
have
on-premises
friends
as
well,
so
it
is
it
fair
to
say
that's
pretty
much.
Nothing
about
this
capabilities
will
be
in
on-premises
and
future.
So.
A
That's
a
great
question,
so
I
don't
think:
we've
totally
decided
that,
but
I
will
say
for
folks
that
have
hybrid
configurations.
You
know.
That's
would
be
my
guidance,
sir
of
where
to
go
if
you're
totally
on
premises.
I,
don't
know
that
I
have
a
good
answer
for
you
right
now.
You
should
get
a
little
bit
more
about
it,
but
the
from
a
hybrid
configuration.
The
wonderful
thing
is
here
with
Cortex's,
as
you
all
know
is
working
on
Microsoft
search
before,
and
so
you
know
the
same
kinds
of
connections
that
we
have.
A
You
think
about
the
the
way
that
we're
able
to
have
that
hybrid
information
come
to
the
Microsoft
graph.
So
if
you
have
on-premise
servers,
you've
got
it
connected
with
us
and
say
you're.
You
know
having
that
information
up
in
the
graph
once
it's
there
and
it's
in
the
Microsoft
graph,
then
it's
available
for
project
cortex
to
basically
crunch
on
and
to
do
discovery
on.
So,
if
you
still
have
you
know
a
bunch
of
stuff,
that's
connected,
obviously,
in
a
hybrid
environment,
still
definitely
part
of
what
we
can
do
with
cortex.
B
But
this
is
a
good
example
of
a
cloud
first
and
we've
been
having
a
cloud
first
solution
for
sharepoint
for
a
quite
a
long
time
already
so
on-premises
sharepoint
2019,
which
is
actually
quite
old,
comparing
what
we
have
in
the
cloud.
But
now
one
final
thing
we're
running
over
stay
out
of
time
as
well.
But
what
about
the
development
side
of
the
house-
and
there
was
a
question
related
on
a
BIS
and
automation?
Is
there
a
way?
B
A
Absolutely
I
think
about
this
is
really
being
a
great
foundation
for
many
new
partner
offerings
and
I
hope
to
see
lots
of
creativity
with
that
as
well.
So
there's
a
couple
of
different
spaces.
There
is
also
a
content
services
partner
program.
If
you
haven't
seen
that
I
think
there's
a
link
out
on
the
tech
community
to
it
as
well.
So
if
you
look
for
content,
Microsoft
content
services
partner
program,
so
we
do,
we
do
have
that
as
well,
so
we're
in
this
area
kind
of
with
content
services,
a
couple
different
things.
A
So
if
I
was
thinking
about
myself
as
a
developer
and
I
was
thinking
about
the
api's,
that
I
would
have
access
to
we're
gonna
do
a
whole
bunch
of
different
things.
One
is
to
have
a
set
of
taxonomy.
Api
is
essentially
for
the
managed
metadata
service.
So
thinking
about
you
know,
we
talked
about
that
import/export
sort
of
functionality.
Obviously
we
want
you
to
be
able
to
automate
that
we
also
want
to
have
different
ways
for
you
to
connect
in
you
know.
A
Potentially
other
term
sets
other
things
that
you
want
to
manage
as
part
of
that
too,
and
then
thinking
about
maybe
different
ways
that
you
might
give
a
different
interface
or
you
know,
build
it
be
able
to
build
on
top
of
the
term,
sets
themselves
so
think
about
managed
metadata
services.
How
do
I
bring
that
maybe
at
other
applications
that
I'm
building
those
kinds
of
areas,
so
that's
kind
of
the
the
MMS
side
of
the
house
now
on
the
sort
of
content,
understanding
side
of
the
house,
obviously
we're
using
a
lot
of
the
azure
capabilities.
A
So
you
know
go
bone
up
on
the
azure
cognitive
services,
because
those
api's
are
ones
that
you
belt
a
call
as
well
and
be
able
to
some
bring
that,
together
with
the
data
that
we
have
inside
of
cortex.
So
there's
a
great
way
to
be
able
to
extract
that
and
get
the
information
you
need
there
and
then
I
think
about
it
on
the
sort
of
the
systems
integrator
side
of
the
house.
A
So
you
know,
there's
a
great
sort
of
different
ways
to
think
about
how
we
do
adoption
for
customers
here
and
then
how
you
build
new
services
on
top
of
capabilities
like
project
cortex.
You
know
it's,
it's
a
great
thing
to
be
able
to
have
this
new
platform,
essentially
inside
of
Microsoft
365,
so
I
always
think
about
it
as
kind
of
a
different
layer.
On
top
of
that,
so
we
will
have
api's
available
for
extensions.
We
will
think
about
the
ways
that
we
do
web
parts.
A
The
traditional
ways
that
we
already
build
with
SPF,
X's
and
people
well
know
thinking
about
the
web
parts
that
you'll
be
able
to
extend
there
and
then
we're
looking
at
different
ways
to
do
extensibility
to
be
able
to
call
both
knowledge
mining
services,
as
well
as
the
the
way
that
we
call
topics
as
well
inside
of
an
organization,
so
lots
of
potential
there
and
we're
definitely
considering
it.
You
know
how
we
make
partners
a
big
part
of
this
as
well.
B
Excellent
that
I
think
everybody,
especially
on
their
side
of
the
house.
People
want
to
hear
about
that
one
and
it's
a
great
great
thing
to
know
that
from
a
day
one
we've
been
thinking
about
the
extensibility
side
of
the
house
as
well.
Now
I
think
that's
it
for
today.
So
thank
you
Eric.
Thank
you.
Nichkhun.
Thank
you
know
me
a
really
great
presentation
and
recapping
actually
what's
what's
coming
and
what
are
the
capabilities,
and
obviously
this
is
one
step
at
a
time
journey.
It's
not
gonna,
be
a
big
bang
and
everybody.
B
Everything
is
gonna,
be
available
immediately
and
right
now
me,
so
we
will
get
some
of
the
stuff
first,
whenever
it's
time
and
then
we'll
reiterate
and
reiterate,
based
on
the
feedback
from
our
users
and
partners,
whether
we
knew
what
needs
to
be
there.
So
it's
gonna
be
a
journey,
but
more
and
more
stuff
will
be
available
later
this
spring
and
SPC
SharePoint
conference
probably
is
gonna,
be
the
big.
Let's
say
some
sort
of
a
big
announcements
probably
is
happening
there.
We
can
assume.