The American Lawyer - Litigation Fall 2011

E-DISCOVERY
The Electronic Eye

WILL COMPUTERS REPLACE LAWYERS IN DOCUMENT REVIEW?

BY VICTOR LI

AuTomATEd E-discovEry
is suddenly hot. Just witness the
frenzy of sniping among com-
petitors when recommind, inc.,
announced in June that it had
received a patent for the predic-
tive coding automated e-discov-
ery process used in its product,
Axcelerate on-demand. rivals
Equivio inc. and orcaTec LLc
immediately fired back that the
recommind patent only covered
a narrow process and that their
products fell outside the scope of
recommind’s claims.
“recommind’s exclusivity claim
was like saying you received a pat-
ent for potato peeling, when in
fact all you got was a patent for a
particular part of a pretty potato
peeler for peeling pink potatoes,”
wrote Warwick sharp, Equivio’s
vice president of marketing and
business development, on his on-
line blog.
The rivalry underscores the
growing number of options for
law firms looking to use tech-
nology in e-discovery to reduce
billable hours. computers have
already proven that they can beat
world champion chess players
and dominate Jeopardy!. Now
they’re threatening to take over
e-discovery—with potentially
huge implications for law firms
and their clients. “Literally a year
ago, no one was interested in
[predictive coding],” says Arnaud
viviers, chief executive officer of
orcaTec. But now several Am
Law 100 firms have adopted auto-
mated e-discovery: squire, sand-
ers & dempsey is using Equivio’s
Equivio>relevance product, and
Wilmer cutler Pickering Hale and
dorr; morgan, Lewis & Bockius;
and Fulbright & Jaworski are us-
ing recommind’s Axcelerate.
(viviers declined to name the law
firms using orcaTec’s document
decisioning suite.)
Prior e-discovery relied on
humans to do keyword searches.
With automated e-discovery us-
ing predictive coding, attorneys
who are familiar with a case
identify a set of documents they
consider relevant and program
the computer to search for similar
ones. The number of documents
in the sample set determines how
in-depth the review process will
be. software then selects similar
documents, analyzing them on
attributes other than keywords—
including context, word frequen-
cy, and other qualities—and gives
each document a relevancy score
as compared to the sample set.
Attorneys can choose whether
to have the software examine all
documents, or only a statistically
significant sample size, which
saves additional time and money.
Typically, attorneys then conduct
quality control by reviewing the
documents chosen as relevant by
the computer.
THE BiggEsT rEAsoN For
adopting predictive coding is to
save money. “The costs of e-dis-
covery are skyrocketing. it’s time
to use this technology, because
if we don’t get a handle on [the
costs], we’ll lose the ability to do
what we’re needed for,” says da-
vid Kessler, cohead of Fulbright
& Jaworski’s e-discovery and in-
formation governance practice,
who says that recommind’s Ax-
celerate with predictive coding
is the firm’s “core review tool.”
stephanie “Tess” Blair, leader
of morgan, Lewis’s edata prac-
tice, agrees, pointing out that the
data volumes have gone up expo-
nentially every year since 2004.
“There comes a point where no
amount of human effort will get
you through that data in an ef-
ficient way,” says Blair, who esti-