Agentic AI, the Next Big Thing, Still Unknown to Many Lawyers

AI agents, representing the next generation of AI, are billed as a major emerging tool to automate routine legal tasks. But a Bloomberg Law survey shows that few lawyers have even heard of them.

More than half of those surveyed in Bloomberg Law’s State of Practice 2025 survey said that they didn’t know what AI agents were, and another 31% said they knew about the technology but hadn’t used it. Only about 5% said they had used AI agents “in a professional setting or to perform a work task.” In-house counsel were three times more likely to have used the technology than were law firm attorneys—12% compared to 4%.

Respondents to the survey included more than 750 in-house counsel and attorneys at law firms.

AI agents are software tools that can work independently to carry out tasks assigned to them by a user. About 80% of senior executives in US companies say AI agents are being used in their companies, according to a May 2025 survey by PwC.

Lawyers might have some catching up to do.

Asked in a follow-up question about the variety of legal tasks that AI agents could perform, nearly 32% of respondents said they wouldn’t trust them to responsibly handle any tasks in the next year.

But 39% said they would use the technology to summarize legal narratives, and 32% said the agents could be used to draft contracts. Attorneys were comfortable using AI agents to work on general legal research about 32% of the time.

Trust Issues

“We are a very conservative profession by trade, and people pay us for our perspective, which we feel is very much of a human interaction,” said Vikas Srinath, a partner at Prospera Law, a firm in Los Angeles. “The idea that any agent could come in with a predefined set of rules and guardrails and effectively mimic 80% to 90% of our process would inherently kind of undercut the advice that we give. That’s at least the psychology I see around me, and it’s a very real psychology.”

Srinath, who responded to the Bloomberg Law survey, is comfortable using agentic AI for research and for supplementing the conversations with clients, he said.

Benjamin de Seingalt, corporate counsel and director of privacy and compliance at MarketVision Research, said that reports of lawyers being disciplined for citing fake cases or other AI hallucinations havent’ helped.

“Even as hallucination becomes somewhat less of an issue, it’s still kind of imprinted on people’s minds as, like, ‘You can’t trust anything from AI,’” de Seingalt, another survey responder said.

There is also confusion on what exactly agentic AI is and what it can do, he said.

“People are envisioning sort of the most automated, you know, the most agentic version of an agent, if you will,” he said.