Meeting & Conference has become a “go to ” destination of
legal start-up executives who act as
“roving evangelists” offering elevator
pitches and a quick demo in hotel lobbies and coffee shops.
where I was about to moderate a panel on AI entitled “Can Robots Be Lawyers? “ Pablo Arredondo, the VP of Research at
Casetext offered to show me a new
product called CARA. After a brief demo,
I was interested enough to schedule a follow up webex.
trained in traditional legal research tools
and analytical techniques – I love the “a-ha”
moment when I realize that a new product
has uncovered insights into data that
was essentially “hiding in plain sight.” I had one of those moments in the CARA demo.
free caselaw research, access to analytical memos from 150 law firms and a “crowd
sourced “citator called “We-cite” is launching their first “fee based” product:CARA. Casetext
currently includes over 7 million primary documents including case law from all 50
states. CARA stands for Computer Assisted Research Assistant.
|Drag and Drop document here|
The Brief as Query.
CARA is very easy to use. All a lawyer has to do is drag and drop a brief into
CARA. Instead of searching with keywords, CARA uses the entire brief as the query
. CARA data-mines the brief by
extracting both the text and the citations from the document. The CARA analysis looks at
direct relationships and “implied” relationships between the cited cases in the
brief and related opinions in the Casetext
database. CARA uses latent semantic analysis
to sort the results. The CARA results report includes a list of “suggested
cases” and a CARA generated analysis of those cases.
Cases – identifies cases not included in your document, which should be
reviewed for relevance before finalizing a brief. A concise summary includes a citation and the
case holding derived from judicial “explanatory parentheticals.” There is a
direct link to the full text of the opinion on Casetext. If a lawyer is
analyzing an opponen’ts brief – this feature could identify vulnerabilities.
Cited Passage– identifies the most cited text from other opinions.
– provides links related law firm memos
discussing the case
|CARA results screen|
Arredondo emphasized that Casetext understands the security
concerns of law firms. Casetext does not retain any brief – extracted data is
deleted after CARA delivers a report.
vocabulary of legal research
is delightfully rich in new research concepts. Arredondo describes the text and citation analysis as a citation fingerprint within the
document which analyzes how much that case discussed and assesses “depth of treatment.”
parentheticals” as the functional equivalents of case summaries written by Westlaw editors. An explanatory
paranthentical is a concise statement of a cited cases holding. The citing judge
signals the explanatory parenthetical with words such as “finding that, holding
that, concurring that.”CARA is the first system to leverage judge generated summaries, which are ubiquitous
throughout the common law. So far
Casetext has identified 2.4 million parenthetical summaries. Arredondo anticipates
that the company will create a searchable
database of explanatory parentheticals which he describes as a “celestial treatise footnote.”
Arredondo’s article “Harvesting and Utilzing Explanatory Parentheticals
“ 1 LEGAL INFO. REV. 31 (2015-2016) which was just published.
powerhouses are already running trials of CARA. CARA is positioned to
supplement rather than replace the major research services (Lexis, Westlaw,
BloombergBNA). Arredondo sees CARA as enhancing the speed of a final review of
a brief by quickly identifying cases that a lawyer may have missed. CARA will also continue to be enhanced with
new features. A jurisdictional filter is expected to be released in the next few weeks.
ThomsonReuters Drafting Assistant will review a brief and check if citations are valid but it does not
explicitly direct a lawyer to cases they may have missed.
precedential behavior using historical
citation trends (Identifying which cases judges tend to cite to when they rule
in favor of or against a motion.) While Ravel provides insights driven by
judges, CARA provides insights driven by briefs. Ravel can be described as “judge
sourcing.” CARA can be described as “brief sourcing” their insights.
Northern District of California, they said that in drafting opinions they invariably reference decisions that were absent from both parties briefs.