Law firm Ropes & Gray is building a proprietary AI tool with OpenAI to do the first pass of transactional diligence, the document-heavy check a buyer runs before closing an acquisition. If it works as described, the grind of reading contracts shrinks and the lawyers' time goes to judgment calls.
The tool combs through thousands of pages, picks out relevant documents and groups them by theme and issue. It also writes summaries and a finished report in the firm's own format. It inventories and categorizes a whole data room, so it looks at relationships between documents rather than each one alone.
"Our clients don't pay us to review contracts," said Howard Glazer, partner and head of Practice AI at the firm. "They pay us to make judgments based on what's in the contracts."
How the firm tested it
Work so far has focused on material commercial contracts. Lawyers compared test outputs with human-written diligence reports from completed deals, then reworked the rubric and report format based on where the drafts fell short. The firm says the tool follows a consistent framework that mirrors the steps its team would take, from initial review to summaries to higher-level review, and that the output is comparable to a capable junior-to-mid-level associate. That comparison is the firm's own, and the post gives no accuracy figures.
Humans stay in the loop
The firm is clear that people remain on the hook. "Our lawyers still need to review everything," said Lichtenstein, whose full name and title the post excerpt doesn't give. Anything delivered to a client must be something the firm can stand behind "100%," so the tool lets lawyers inspect interim summaries and the source documents.
Ropes & Gray has publicly described an ambition to produce a detailed, issue-level report in hours rather than weeks. That is a goal, not a measured result. The post also mentions GPT-6 Astra alongside the in-house tool as part of how the firm solves client problems faster, but it doesn't say how the two fit together.
What we don't know
The post doesn't say when the tool will be in wide use, how many deals have run through it, how often it misses documents, or how clients will be billed for faster work. It is also an OpenAI-published case study, so the praise comes from the partners. This story rests on that post alone: no outside law firm, client or legal tech analyst has confirmed that the time savings hold up on live deals, and Ropes & Gray has not given us details beyond what OpenAI published.
Watch for hard numbers on error rates and turnaround once real deals go through. If a first draft arrives in hours, should clients expect to pay less for diligence, or just get more of it?
Sources
Reported by the WattsUpNext desk from the sources linked below. Spot an error? Tell us at corrections@wattsupnext.com.
The WattsUpNext Brief
Get stories like this in your inbox.
One email each weekday, only the topics you choose. Real news, sourced, no fluff. Unsubscribe in one click.
