It
was
the
second
non-IP
request
of
the
week
from
a
longtime
client,
made
through
their
new
in-house
counsel.
The
first
of
the
two
requests
was
easy
enough
to
dispatch,
even
though
it
involved
an
already-filed
litigation.
The
client
and
I
had
gotten
a
similar
case
dismissed
a
few
years
back,
even
before
answering
the
complaint.
All
I
needed
to
do
was
go
back
through
my
emails
around
that
dispute.
Once
I
refreshed
my
recollection
on
the
investigation
we
had
done
in
that
earlier
case,
I
was
able
to
guide
my
client’s
new
in-house
person
as
to
what
information
he
needed
from
the
business
unit
and
a
relevant
outside
vendor.
With
that
info
in
hand,
it
was
then
a
simple
matter
of
presenting
it
to
opposing
counsel
and
securing
their
agreement
to
drop
the
case.
This
time,
we
didn’t
even
need
to
engage
Biglaw
co-counsel
with
specific
expertise
in
this
type
of
litigation,
as
we
did
the
first
time
around.
Score
one
for
the
value
of
long-term
client-attorney
relationships
and
institutional
knowledge.
The
second
matter
that
arose,
however,
was
a
bit
more
challenging.
To
start,
the
client
had
received
a
notice
of
the
potential
claim
a
while
back,
with
a
limited
time
remaining
to
reply.
Complicating
matters,
the
likelihood
of
the
client
getting
indemnified
by
its
supplier
was
nonexistent,
since
the
product
had
been
discontinued
and
recent
sales
were
just
an
attempt
to
clear
out
remaining
inventory.
Moreover,
the
plaintiff’s
attorney
was
threatening
to
file
a
lawsuit
in
state
court,
across
the
country
from
where
my
client
is
located,
which
would
necessitate
me
helping
the
client
identify
competent
local
counsel
in
an
area
of
law
I
was
unfamiliar
with.
Gulping
down
this
witches’
brew
was
an
unattractive
proposition
on
my
part,
especially
because
the
client
was
asking
for
immediate
help
preparing
for
a
last-ditch
effort
to
get
the
matter
resolved
prelawsuit
via
a
scheduled
call
with
plaintiff’s
counsel.
Because
the
matter
was
time-sensitive,
I
did
a
quick
read
of
opposing
counsel’s
demand
letter
and
followed
up
with
my
client
to
confirm
a
factual
point
the
letter
raised.
Because
of
my
unfamiliarity
with
the
type
of
claim,
my
plan
was
to
suggest
to
the
client
that
they
ask
around
for
recommendations
to
a
firm
based
in
the
state
where
the
dispute
was
centered.
In
fact,
that
was
my
initial
recommendation
to
the
client
by
email.
To
the
client’s
credit,
they
pressed
for
more
guidance,
which
prompted
me
to
see
what
AI
could
do
to
help
direct
me
to
next
steps.
So
I
fired
up
one
of
my
trusty
AI
tools
and
started
by
asking
for
a
primer
on
the
type
of
claim
being
raised
against
my
client.
To
say
it
was
nowhere
as
complex
as
an
IP
dispute
would
be
an
understatement.
My
next
step
was
to
ask
the
AI
what
would
be
the
best
approach
to
“head
off”
a
filing
against
my
client.
The
AI
returned
a
few
different
strategies,
including
some
that
my
client
could
not
execute
on,
(e.g.,
qualifying
for
a
safe
harbor
based
on
the
size
of
the
business,
as
one
example). I
knew
all
along
that
I
would
need
to
upload
the
demand
letter
my
client
had
received
to
get
more
tailored
advice
—
but
I
felt
that
it
was
important
to
get
some
more
general
guidance
from
the
AI
first,
so
that
I
could
double
check
that
it
was
giving
me
consistent
responses.
As
a
next
step,
I
uploaded
the
demand
letter
from
opposing
counsel.
The
AI
analyzed
it
in
seconds
—
and
even
pulled
the
underlying
notice
that
opposing
counsel
had
filed
with
the
state
authority
—
before
returning
a
detailed,
step-by-step
approach
to
counseling
the
client
on
their
legal
options.
For
step
one,
I
needed
to
confirm
how
many
units
had
been
sold
in
the
relevant
period.
A
quick
email
exchange
with
the
client
later,
and
I
had
the
information.
Armed
with
that
data,
the
AI
was
able
to
prepare
a
more
detailed
approach
for
my
client
to
take
on
the
call
with
opposing
counsel.
It
even
suggested
an
opening
settlement
offer,
large
enough
to
force
opposing
counsel’s
hand
toward
a
reasonable
response,
but
low
enough
to
reflect
the
limited
exposure
my
client
faced
even
if
a
case
was
filed.
I
then
switched
gears,
prompting
the
AI
to
give
me
my
client’s
maximum
exposure
under
the
law,
taking
all
facts
in
the
favor
of
opposing
counsel
as
set
forth
in
their
letter.
It
confirmed
that
the
biggest
risk
to
my
client
was
getting
entangled
in
an
expensive
legal
proceeding,
where
even
a
narrow
loss
on
the
merits
could
result
in
a
stinging
attorney’s
fees
award.
With
all
that
analysis
in
hand,
coupled
with
my
own
verification
of
everything
the
AI
had
suggested,
I
then
felt
comfortable
advising
the
client
on
how
to
approach
the
initial
settlement
discussion.
I
reckoned
that
even
were
that
talk
proven
unsuccessful,
we
would
have
learned
something
about
the
other
side
by
virtue
of
the
discussion,
with
the
ability
to
pivot
to
the
original
strategy
of
hiring
competent
local
defense
counsel
as
needed.
And
if
our
AI-assisted
approach
proves
to
work,
then
the
client —
even
after
paying
me
for
my
counsel
—
would
have
saved
significant
money
in
outside
counsel
fees,
while
avoiding
an
ongoing
business
distraction
with
nothing
to
gain.
Moreover,
because
the
client
has
traditionally
worked
with
me
on
a
flat
fee
per
project
basis,
I
didn’t
have
to
worry
too
much
about
how
much
time
using
AI
may
(or
may
not)
have
saved
me.
Instead,
both
the
client
and
I
would
have
benefited
by
the
research
capabilities
of
the
AI,
as
well
as
the
speed
with
which
it
suggested
different
approaches
that
I
was
allowed
to
pressure
test.
Only
on
that
foundation,
as
well
as
my
confirmation
with
independent
research
of
everything
it
pointed
me
to,
did
my
legal
judgment
kick
in
to
allow
me
to
advise
the
client
in
the
way
I
felt
best.
Ultimately,
I
would
couch
this
experience
as
a
successful
use
of
AI,
irrespective
of
whether
this
dispute
resolved
on
a
faster
or
slower
track.
Perhaps
what
is
most
interesting
to
me
about
the
experience
is
the
fact
that
this
was
not
a
project
that
I
would
have
felt
comfortable
delegating
to
an
associate,
whether
or
not
they
used
AI.
If
they
didn’t
use
AI,
I
wouldn’t
expect
an
IP
associate
to
give
me
accurate
guidance
on
an
area
of
law
unfamiliar
to
us
both.
And
if
they
did
use
AI,
having
them
regurgitate
for
me
the
AI’s
output
would
have
been
of
little
additional
help
—
and
likely
more
inefficient
with
a
higher
chance
of
error.
Put
another
way,
just
as
I
would
not
substitute
an
associate’s
judgment
for
my
own,
neither
would
I
outsource
my
judgment
to
any
AI
tool.
Going
forward,
I
will
look
for
an
example
from
my
practice
where
handing
off
an
AI-assisted
assignment
an
associate’s
way
would
prove
to
have
been
a
better
approach.
In
the
meantime,
using
AI
on
this
one
project
was
all
I
could
ask
for.
Please
feel
free
to
send
comments
or
questions
to
me
at
[email protected]
or
via
Twitter:
@gkroub.
Any
topic
suggestions
or
thoughts
are
most
welcome.
Gaston
Kroub
lives
in
Brooklyn
and
is
a
founding
partner
of
K2K
IP
Law,
an
intellectual
property
litigation
boutique
that
also
serves
as
a
leading
consultancy
on
patent
issues
for
the
investment
community.
Gaston’s
practice
focuses
on
intellectual
property
litigation
and
related
counseling,
with
a
strong
focus
on
patent
matters.
You
can
reach
him
at
[email protected]
or
follow
him
on
Twitter:
@gkroub.