One
of
the
key
things
I
learned
from
a
number
of
mentors
in
the
mass
tort
defense
league
was
the
importance
of
front-end
loading
cases.
By
that
I
mean
getting
a
grasp
early
on
of
the
exposure
and
the
“oh
shit”
documents
and
then
immediately
formulating
the
appropriate
case
strategy
based
on
that
grasp.
Over
and
over
again,
the
approach
proved
itself
as
the
case
progressed.
Another
thing
I
learned
along
the
way
was
the
improved
depositions
and
cross
examination
of
witnesses
where
I,
as
the
most
knowledgeable
and
experienced
lawyer
on
the
case,
could
myself
review
and
locate
the
most
important
documents,
as
opposed
to
asking
an
associate
or
paralegal
to
do
that
work
and
then
report
to
me.
The
problem
with
both,
of
course,
was
always
the
cost
and
time
involved
to
do
this
important
work.
All
too
often
clients
would
balk
at
paying
for
this
work
early
on.
Convincing
them
otherwise
was
a
pain,
to
say
the
least.
The
Deep
Dive
Tool
Enter
Everlaw’s
new
Deep
Dive
tool,
which
promises
to
address
both
challenges.
The
tool
is
an
enhancement
of
what
was
previously
called
Project
Query,
according
to
the
Everlaw
press
release.
Deep
Dive
was
announced
in
open
beta
at
the
ILTA
conference
this
week.
In
essence,
the
tool
allows
the
user
to
query
a
large
data
set
of
inputted
documents
using
natural
language.
The
tool
allows
users
to
simply
ask
questions
related
to
specific
issues,
parties,
or
events
and
get
answers.
In
a
class
action,
you
might
ask
the
tool
something
like
“is
there
any
support
for
the
claims
that
there
are
common
questions
of
fact
among
the
various
putative
class
members.”
The
answers
are
supported
with
a
list
of
facts
and
referenceable
documents
so
users
can
dive
deeper
into
the
breakdown
of
information
available
as
part
of
the
response.
In
other
words,
the
tool
cites
the
documents
that
it
uses
to
answer
the
questions.
You
can
then
drill
down
into
the
documents
and
ask
more
specific
questions
based
on
a
subset
of
the
cited
documents.
The
tool
does
not
go
outside
the
documents
to
answer
any
inquiry.
Importantly,
according
to
AJ
Shankar,
Everlaw’s founder,
who
demoed
the
product
for
a
group
of
journalists
over
lunch,
if
Deep
Dive
can’t
find
any
documents
to
respond
to
the
question,
it
will
say
so.
(In
the
demo,
based
on
the
question
one
of
us
asked
as
to
who
the
best
legal
tech
journalist
was,
Deep
Dive
demurred
and
said
there
were
no
documents
in
its
data
base
to
answer
the
questions,
thankfully.)
The
demo
certainly
backed
up
the
above
claims
from
Everlaw’s
press
release.
According
to
Chuck
Kellner,
Everlaw
Advisor,
who
also
attended
the
demo,
Everlaw
envisioned
three
use
cases:
1.
Undertaking
a
data
dump
before
depositions
to
help
either
the
taking
or
defending
lawyer
better
prepare.
2.
Using
the
tool
at
the
beginning
of
case
to
assess
exposure
etc.
without
a
huge
investment
of
time
and
cost
and
before
discovery.
3.
Designing
document
review
and
coding
projects.
Use
of
the
tool
would
eliminate
the
“I
don’t
know
what
to
look
for”
problem.
The
tool
can
give
an
overview
of
the
documents
to
help
design
the
review
and
coding.
The
first
two
of
these
were
particularly
music
to
my
ears.
Kellner
confirmed
what
I
was
thinking.
It
is
designed
not
for
those
working
on
a
case
but
for
the
case
lead
and
strategist.
And
as
I
said
before,
these
are
the
people
that
know
the
overall
case
the
best
and
know
the
client
the
best.
Giving
them
the
ability
to
ask
questions
will
inevitability
lead
to
better
results.
All
too
often,
eDiscovery
vendors
portray
what
their
tools
will
do
defensively
and
not
offensively
in
the
hands
of
an
experienced
lawyer.
And
the
ability
to
ask
what
documents
might
support
the
other
side’s
allegation
is
golden
to
a
front-end
loading
and
strategy
process.
Simply
plug
in
the
an
allegation and
then
ask
what
supports
or
does
not
support
that
claim.
A
Credible
Demo
But,
here’s
another
lesson
from
the
Everlaw
demonstration.
In
a
conference
where
too
many
vendors
are
promising
the
sun,
the
moon,
and
the
stars,
Shankar
was
refreshingly
honest
about
what
the
tool
can
and
can’t
do.
He
recognized
that
the
questions
lawyers
need
answers
to
aren’t
necessarily
easy
for
AI
tools
to
answer.
That
Deep
Dive
might
make
mistakes,
that
it’s
an
aid,
not
a
panacea.
Shankar
even
did
the
demo
live,
not
using
a
screen
video
like
most
use
in
vendor
presentations.
The
technology
didn’t
necessarily
work
perfectly
during
the
demo
(those
who
do
this
in
real
life
know
this
is
an
issue
for
a
variety
of
reasons
that
have
nothing
to
do
with
the
product).
He
let
the
audience
ask
questions
of
the
tool.
He
admitted
that
while
he
thought
he
knew
what
the
tool
would
do
in
the
demo,
it
wasn’t
foolproof.
“It
can’t
do
everything,”
he
cautioned.
The
approach
was
refreshing
and
enhanced
both
his
and
Everlaw’s
credibility
in
my
mind.
A
useful
tool
and
a
credible
demo.
Imagine
that.
Stephen
Embry
is
a
lawyer,
speaker,
blogger,
and
writer.
He
publishes TechLaw
Crossroads,
a
blog
devoted
to
the
examination
of
the
tension
between
technology,
the
law,
and
the
practice
of
law.