When AI is your law-clerk assistant

Raul Gastesi, founding partner at law firm Gastesi Lopez Mestre & Cobiella, was up in the air—and so was the lawsuit he was trying to file.

Just over a year ago, while aboard a flight, Gastesi had everything he needed to write that lawsuit, including attachments, invoices, and demand letters. Instead of emailing them off to a junior clerk, though, he turned to the LexisNexis AI platform.

The prompt went a little something like: I want to file a lawsuit in Broward County Circuit Court. See the attached letter with the invoices.

“And boom, it printed up a complaint,” Gastesi told us.

The output did need some work, he mentioned, like formatting and fact checking, but the tool saved him time and effort. A year later, the law firm has integrated AI into many of its workflows, including LexisNexis for research and initial drafts and Anthropic’s Claude for formatting and synthesizing.

AI tools such as lawsuit generators have sped up lawyers’ legal work. But the speed has introduced the need for best practices and safeguards. Mainly: check your work!

All rise! A Thomson Reuters report, released in August 2025, found that legal professionals are turning to AI for a number of tasks, including document review (77%), legal research (74%), and document summarization (74%).

AI company Harvey, founded in 2022, built a platform that provides insights specifically for lawyers and in-house legal teams. Lawyers can ask the tool for help with tasks like drafting clauses, reviewing portfolios, analyzing contracts, and following regulations.

John LaBarre, general counsel at Harvey, has even heard of litigators using the platform as they prepare to depose an opposing side’s expert. A litigator might feed Harvey their report, ask it to find any other works from the author, and then note any inconsistencies between past and present statements.

“What’s better than catching your expert on an inconsistency?…That’s a great moment for a litigator, and AI can help you find those rather quickly,” he said.

LaBarre sees AI impacting lawyers in similar ways to how the technology changed software developers, making the latter “project managers” rather than pure coders. He sees lawyers already orchestrating multiple activities (through agent-driven workflows) and having to review outputs as they arrive.

Disorder in the court. Manually reviewing those outputs is an essential step for an AI-enabled lawyer, as AI hallucinations have made embarrassing appearances in court. A judge in Mississippi in June reportedly disqualified lawyers following AI-fabricated legal citations.

One way to minimize hallucinations, LaBarre advised, is to use AI tools that cite the sources of their statements.

That check-your-work attitude is the same on Gastesi’s team. Christopher J. Mangos, an associate attorney at Gastesi’s firm, sees the AI as “basically a law clerk or an assistant”; he’s ultimately responsible for its actions, including possible hallucinations.

“It’s about being on top of review and analysis of everything, and telling it: ‘Be more specific, verify, show me where you got this,” Mangos said.

After review…Years before mainstream AI usage at law firms, Craig A. Greening, CEO and head litigator at the Greening Law Group, handled a case from the FBI involving “probably a million pages of evidence.” That kind of reading required hiring part-timers and having them sign non-disclosure agreements to summarize the documents.

Now he uses transcription tools like Rev, which allow him to explore lots of documentation, including video. Greening can collect feeds from dash cams and police-officer body cams, along with relevant documents like traffic-stop summaries and find inconsistencies in statements, asking specific questions like, What was the traffic stop for? Why did the officer arrest the person? What evidence did they find?

“I think that the real risk isn’t AI replacing defense attorneys. I think the risk is a defense lawyer without these tools will probably miss the exculpatory moment that’s sitting in their own discovery,” he said.

The impact of an AI researcher on a human researcher. The true answer about how GenAI may impact jobs, revenue, or the profession at large “may be years—if not decades—away,” according to Thomson Reuters’s 2025 report, which also cited that 15% of surveyed legal professionals see GenAI as a “major threat” (and 44% see “somewhat of a threat”) in terms of impact on jobs.

In the meantime, Gastesi sees AI boosting legal professionals’ research: “I find that the work is coming back to me more complete with more analysis. I see an improvement in the quality of the work that is coming back to me,” he said.