Ray Mangum Featured in Relativity’s “Understanding the Promise and Limitations of Agentic AI for Legal” Toolkit

Ray Mangum recently contributed insights to Relativity’s “Understanding the Promise and Limitations of Agentic AI for Legal” toolkit, an in-depth resource that examines the evolving role of agentic AI in the legal field.

This comprehensive toolkit provides legal professionals with clear definitions, ethical considerations, practical use cases, and guidance on how to responsibly integrate agentic AI into legal workflows while maintaining human oversight. 

Ray’s key contributions to this toolkit are outlined below.

. . .

What is Agentic AI?

“If you see this question, ‘what is agentic AI,’ and you feel a little uncertain— you were just getting your head around generative AI, and now we’re already on agentic AI—well, you are in very, very good company,” Ray Mangum, partner at Redgrave, says.  

“We’re all trying to get our heads around this.  If you follow the news around AI, you’ll see lots of discussion: it’s a bit of a running joke, asking ‘what is an agent or agentic AI?’”  he continues.  “That said, I don’t think it’s a meaningless term.”  

Mangum notes that he’s involved in a working group with the Sedona Conference, and part of their mission is to help set some formal definitions of these terms for the legal technology space.  We’re all eager to see them; you can follow along on their page.

In the meantime, most definitions share some similarities: agentic AI does something on your behalf.  It doesn’t only answer questions; it takes some action according to your instructions.

Put simply, agentic AI encompasses systems that perceive, think, and act on their own to meet user-set goals.

. . .

A Word on Safe Sandboxes

Ray Mangum puts it simply: “Playing with it is the best way to learn.  Go out and do it and don’t be intimidated—just start using it.  Come up with a real problem and try to solve that problem.  Even if you don’t make it all the way through, you’re going to learn a tremendous amount from doing this.”  

The best way to build intuition around agentic AI is to play—but to play safely.

Learn more about agentic AI and the safety considerations surrounding it by accessing Relativity’s toolkit here.