Agentic AI in Banking: Enthusiasm Outpaces Production Deployment

From Hype to Hard Questions

Agentic AI has become one of the most discussed technologies in banking and financial services. The open question, according to a new expert roundtable hosted by GlobalData, is how much of that enthusiasm is translating into real enterprise impact.

The session, titled “Agentic AI in Banking & Financial Services: Beyond the Hype,” brought together Jeff Veis, Chief Marketing Officer at Impetus Technologies; Deepak Khosla, Chief Growth Officer and Head of AI Business at the same firm; and Stephen Walker, Retail Banking Analyst at GlobalData.

Why Scale Remains Elusive

The discussion moved quickly past industry promotion to examine a persistent gap: many financial institutions are experimenting with agentic AI, yet relatively few have deployed it at scale.

The participants identified several barriers to production adoption, including lack of data readiness, fragmented enterprise context, governance gaps, safety risks and regulatory expectations. Each of these constraints weighs differently in finance than in other sectors, where the consequences of an underperforming model are often less severe.

The Higher Bar in Financial Services

Khosla framed the distinction sharply. “An agent in banking is not just summarizing a document,” he stated at the event. “It could influence and impact credit, fraud, payments, customer treatment, reporting, or advice. The bar for production is therefore much higher in the banking and financial services sector.”

That elevated threshold explains why experimentation has not translated automatically into deployment. The functions an agent might touch carry operational, financial and reputational consequences that demand more than a working prototype.

Agentic AI adoption in banking depends less on model capability than on the quality of enterprise context available to agents.

Context as the Foundation

Khosla emphasised that successful adoption depends on building strong AI-ready data foundations and grounding AI agents and systems in proprietary business processes, operational realities, historical interactions and governance policies.

“Our approach starts with the belief that agentic AI success depends on the quality of enterprise context available to agents,” he said. “If that context is fragmented, stale or poorly governed, they will fail in production.”

The point reframes the adoption problem. The limiting factor is not necessarily the agent itself, but the reliability of the information environment in which it operates.

Engineering the Enterprise Context

The session concluded that financial institutions should focus on engineering enterprise context by building trusted semantic layers, ontologies and knowledge graphs that AI agents can reliably reason over.

That groundwork, the participants argued, is what enables institutions to power strategic high-impact use cases where returns are measurable and risk can be managed. The emphasis falls on use cases that can be evaluated, not on breadth of deployment.

What the Roundtable Signals

For banks weighing agentic AI, the discussion suggests a sequencing question rather than a purely technological one. Data readiness, governance and regulatory expectations are presented as prerequisites, not afterthoughts.

The roundtable was produced as a GlobalData expert session, and the full discussion is available in the recorded format referenced by the organisers.