Soon, we’ll be chatting about a new buzzword: “agentic AI”

Like any hot technology, the rapidly evolving field of artificial intelligence has its own rapidly evolving vocabulary. There’s AGI, or artificial general intelligence, which is basically a software robot smarter than any of us humans. There’s “hyperscalers,” which are basically data center giants that are rich enough to create their own AI.

But if you had to pick the most important AI buzzword for 2025, “agentic AI” could very well be the winner.

Let’s say you’re the head of an investment firm, like Doug Clinton at Intelligent Alpha.

In prehistoric times, like last year, you might tell a junior analyst, “Hey there Billy, can you go through the SEC filings of Company X, figure out the risks to their business, then email them some questions?”

Now in 2024, an AI agent can do a lot of that on its own. And its emails will probably have fewer grammatical errors than Billy’s.

“So the email says, ‘Could you provide me some detail on specific initiatives Company X is undertaking to attract and retain a more diverse consumer base?'” Clinton said. That’s part of a real email an AI agent wrote and sent to a real company for Clinton — he’s withholding the real company’s name.

“I think these agents have been pretty good, you know. Often, I would say, you know, they look like reports that I would expect to see from a human analyst,” he said.

With AI chatbots like ChatGPT, you can say, plan me a trip to Ireland for a budget-conscious family of four, and it’ll give you a hypothetical agenda.

With AI agents, you say, plan that trip to Ireland and it’ll go ahead and book the discount flights and Guinness brewery tour for you.

“You’re creating software that can actually do things in the real world, multiple steps where you don’t actually have to write all of the code and give it exact directions on what to do,” said Alkesh Shah, head of Americas software research at Bank of America.

Businesses are salivating over agentic AI. Ruchir Puri at IBM said the speed at which AI agents are adopted within a company will depend on how employees react to them.

“It’s about the trust of the system. Until I come to trust the system totally, I’m still in the driver’s seat. I’m driving,” he said.

Puri’s talking about employees trusting that the robot won’t make a mistake — not necessarily trusting that the robot won’t, you know, just replace them.