
Data-driven lending at speed: How Carmoola is using AI to improve decisions and stay ahead
What does it take to modernise credit decisioning in a fast-moving fintech and deliver measurable impact? In this episode of The esynergy Podcast, Grant Ongers speaks with Dan Kellett, Director of Lending and Data Analytics at Carmoola and former UK Chief Data Officer at Capital One, about applying data and AI to real-world lending challenges. After 24 years at Capital One, Dan made the move to a smaller, faster-paced organisation to get closer to delivery and strategic impact. He shares how Carmoola increased loan approvals by 10% while maintaining stable risk, by upgrading credit models with more recent data, richer predictive sources, including enhanced credit reference agency data and open banking and more sophisticated algorithms. The conversation also explores how customer segmentation enables more tailored lending strategies and communications, and how organisations can balance innovation with regulatory expectations. Key topics covered: • Improving credit decisioning with better data and more advanced models • Using open banking and enhanced data sources to increase predictive power • Achieving higher approvals while maintaining risk discipline • The role of customer segmentation in lending and engagement • Insights from the FCA/Bank of England AI forum and evolving governance • A practical, three-layer approach to embedding AI across an organisation • Why leaders must adapt and avoid over-relying on past approaches Dan also shares his perspective on leadership in a rapidly changing landscape, including the need to embrace change, let go of legacy thinking, and build organisations that can move faster and learn continuously. This episode is essential listening for CTOs, data leaders, fintech executives, and AI practitioners looking to deliver measurable business impact through data and intelligent decisioning.





