Inside NatWest's Approach to Agentic AI | Dr. Paul Dongha
In this episode of the ipushpull Capital Markets FinTech Forum podcast, Matthew Cheung sits down with Dr. Paul Dongha, Head of Responsible AI and AI Strategy at NatWest Group, to unpack what agentic AI actually means for banking, and why governance is what decides whether AI scales or stalls.Paul has been thinking about AI agents since before most people had heard the term. He finished his PhD on AI agent reasoning and the alignment problem back in the 1990s, then spent almost three decades putting it to work inside seven banks. Today he leads NatWest's efforts to deploy AI in a way that's safe, transparent, and auditable, and he leads the bank's collaboration with the FCA, where he's live testing agentic AI. Before NatWest he was Group Head of Data and AI Ethics at Lloyds. He's also co-author of Governing the Machine, How to Navigate the Risks of AI and Unlock Its True Potential, an FT book of the month, and he's about to launch a new course called Responsible AI in Practice.In our conversation, we covered:- Why AI governance is the unlock for scaling AI, not a compliance burden- Why LLMs will become commodities, and the browser war that proves itThe levels of autonomy, from human in control to on the loop- The alignment problem he studied in the 1990s, still unsolved today- Using customer data to bring back the bank manager who actually knew you- Why agents should be treated as tools, not given employee logins- MCP as a base layer, and the governance checks still missing- How the EU AI Act and FCA consumer duty fit what firms were doing anyway- Why AI literacy matters most, and taking the hype with a pinch of saltConnect with Dr. Paul Dongha:https://www.linkedin.com/in/paul-dongha/Resources:https://www.amazon.co.uk/Governing-Machine-navigate-unlock-potential/dp/139942629Xhttps://maven.com/georgiana-marsic/ai-governance-framework For more information go to: https://ipushpull.com
