When
generative
AI
first
crashed
into
legal,
my
biggest
question
wasn’t
whether
it
was
useful,
but
if
lawyers
would
embrace
the
idea
of
a
chatbot.
It
worked
for
the
masses,
but
would
lawyers
really
want
to
do
their
work
by
typing
vague
questions
to
an
obsequious
robot?
It
turned
out,
for
most
lawyers,
“typing
vague
questions
to
an
obsequious
robot”
was
how
they’d
seen
junior
associates
all
along
and
the
chatbot
interface
took
off.
This
week,
NetDocuments
unveiled
a
new
feature
that
also
borrows
from
the
consumer
AI
world,
announcing
what
it’s
calling
the
first
legal
context
graph
—
a
continuously
indexed
structure
that
maps
the
relationships
between
every
matter,
document,
and
communication
in
a
firm
that
connects
across
hundreds
of
millions
of
records.
It
should
go
without
saying
that
it
does
this
with
full
respect
for
the
firm’s
existing
permissions
and
ethical
walls.
In
the
consumer
AI
world,
Obsidian
took
off
as
a
note-taking
repository
because
it
allowed
users
to
create
graphs
to
visualize
the
relationships
between
notes.
NetDocuments
offers
that
concept
on
steroids,
building
out
the
graphical
approach
for
AI’s
benefit.
To
quote
from
the
press
release,
the
result
is
not
a
new
interface,
but
a
“fundamental
shift
in
what
the
platform
is.”
Not
that
the
interface
isn’t
getting
a
redesign
too,
a
product
of,
according
to
the
company,
39
design
studies
and
more
than
1,500
participants.
But,
and
for
law
firms
this
is
the
most
important
news,
users
can
toggle
between
the
existing
and
new
platform
with
no
technical
migration.
For
those
lawyers
who
hate
change.
Taking
this
graphical
approach,
NetDocuments
endeavors
to
build
a
system
that
truly
understands
the
work
it’s
storing.
By
constantly
connecting
files
in
the
whole
repository,
a
lawyer
opening
up
an
unfamiliar
matter
can
see
the
full
context,
a
summary,
key
parties,
an
activity
timeline,
and
the
people
who
have
done
this
work
before.
The
AI
working
inside
NetDocuments
(or
external
tools
connected
to
NetDocs
via
MCP,
like
Claude,
ChatGPT,
or
other
legal
AI
applications)
can
now
navigate
all
these
document
relationships
to
deliver
the
overview
and
set
the
stage
for
the
user
to
begin
working
within
the
matter.
The
premise
is
that
bottleneck
in
legal
AI
is
in
what
it
can
see.
Everyone’s
had
an
LLM
get
lazy
and
go
rogue
when
it
can’t
immediately
find
what
it’s
looking
for.
Building
out
these
connections
greases
the
wheels
for
the
AI.
NetDocuments
designed
the
system
with
AWS
and
Elastic
to
continuously
process
and
connect
hundreds
of
millions
of
documents
under
strict
governance.
It
indexes
at
three
levels.
There’s
the
document
level
—
classification,
extracted
entities,
version
history
with
auto-generated
summaries
of
what
changed
and
why.
There’s
the
matter
level
—
auto-assembled
overviews,
key
parties,
dates,
team
visibility,
an
activity
timeline
that
surfaces
both
human
and
AI
agent
activity.
And
there’s
the
global
level
—
natural
language
search
across
the
whole
repository,
governed
by
the
same
access
controls
and
ethical
walls
that
already
exist
in
the
DMS.
A
good
example
of
what
this
looks
like
in
action
is
its
co-authoring
feature
in
Microsoft
Word.
Attorneys
can
issue
natural-language
instructions
to
update
sections
of
a
brief
based
on,
say,
a
newly
added
expert
report.
The
graphical
connections
allow
the
AI
to
understand
the
request
and
go
draw
upon
the
expert
report
it
knows
lives
within
the
system.
It’s
all
about
connections.
The
AI
influencers
out
there
talk
about
maximizing
AI
by
building
a
“second
brain”
with
connected
files.
NetDocuments
has
its
indexers
doing
that
automatically
every
time
something
gets
added.
Lawyers
have
spent
years
organizing
files
with
folders
and
idiosyncratic
file
names.
Lawyers
are
trained
to
think
in
terms
of
v.FINAL_FINAL_clean_jp_v3.
But,
like
the
chatbot
interface
before,
NetDocuments
is
betting
that
lawyers
will
embrace
this
new
mode
of
thinking
about
document
management.
Because
the
AI
will
definitely
prefer
it.
Joe
Patrice is
a
senior
editor
at
Above
the
Law
and
co-host
of
Thinking
Like
A
Lawyer.
Feel
free
to email
any
tips,
questions,
or
comments.
Follow
him
on Twitter or
Bluesky
if
you’re
interested
in
law,
politics,
and
a
healthy
dose
of
college
sports
news.
Joe
also
serves
as
a
Managing
Director
at
RPN
Executive
Search.