How Agentic AI adapts to legacy and leads the way

What if Al didn’t just assist but worked like a team? In finance, Agentic Al is doing just that: teaming up, factchecking, adapting to legacy systems, and transforming workflows without making a sound. The bots aren’t coming. They've already clocked in

- Ashok Pandey [e-mail]

When we spoke with Jon O’Donnell, Chief Operating Officer, Acuity Knowledge Partners, the conversation quickly moved past the buzzwords. No hype, no generic “Al will change everything” fluff. Instead, what unfolded was a grounded, detailed look at how Agentic Al, an emerging architecture of autonomous, collaborative Al agents, is quietly transforming the way financial services firms operate.

“We started with copilots,” O’Donnell explained, “but quickly realized the future wasn’t about single tools. It was about teams, AI agents that collaborate like humans, validate each other's work, and adapt to real-world data complexities.”

And just like that, the frame shifts: from AI as an assistant, to AI as a workforce.

Not just Al: An entire workforce of algorithms

Artificial intelligence is changing. Not in the look, it can summarize a PDF, it let meme behave like a team of analysts, collaborate, flag inconsistencies, and generate outputs at scale. Welcome to the world of Agentic AI. This new approach doesn't rely on one large model answering all your questions. Instead, it mimics how human teams operatewith specialization, coordination, checks, and balances. In sectors like banking, asset management, and private equity, where accuracy and trust are nonnegotiable, this model is starting to reshape how work gets done.

From copilots to digital departments

Earlier generations of enterprise AI systems were linear. A prompt in, a response out. At the best, they behaved like copilots, offering suggestions, finishing sentences, and surfacing data, etc. But today's leading financial services firms are experimenting with something more sophisticated: AI “fleets” made up of semi-autonomous agents, each trained for a specific role.