Abhishek Mittal: How Agentic AI Fights Financial Crime — Even with Imperfect Data

Description

In this episode of CAIO Connect, host Sanjay Puri sits down with Abhishek Mittal, EVP and Chief Product & AI Officer at AML RightSource, for a deep and engaging conversation on how AI is transforming the fight against financial crime. From building AI systems that thrive on imperfect data to pioneering “digital FTEs,” Abhishek shares a pragmatic view of what it takes to make AI work in the real world — where innovation, governance, and human expertise come together to drive meaningful, enterprise-scale change. 💡 5 Key Takeaways: 🔹 AI with Imperfect Data: Why waiting for “perfect data” is a trap — and how to iterate your way to AI success. 🔹 Synthetic Data Evolution: How synthetic data is training smarter models and preparing them for unseen scenarios. 🔹 Financial Crime as a Global Issue: If financial crime were a country, it’d have the world’s 7th largest GDP — and AI could be its toughest opponent. 🔹 Digital FTEs & Agentic AI: How AI agents and human experts are working side by side to redefine efficiency and decision-making. 🔹 Innovation at Scale: Abhishek’s 5-step framework for turning bold ideas into commercial AI success. If you enjoyed this conversation, don’t forget to like, share, and subscribe for more insights from global AI leaders shaping the future of enterprise transformation. Resources Mentioned: https://www.linkedin.com/in/abhishek-mittal-7115348/ Timestamps: 00:00 — Introduction: Financial crime as “7th largest GDP”; podcast begins 00:30 — How AI is transforming anti-money laundering (AML) 01:10 — The ethos of AML RightSource: “Financial crime fighter” mindset 02:45 — Global scale and complexity of financial crime 04:35 — Crypto, digital money & regulatory challenges 07:00 — Elder abuse scams and the role of banks in prevention 09:00 — Education, awareness, and the importance of policy 11:40 — Data doesn’t have to be perfect—how to iterate AI with imperfect data 15:00 — Synthetic data, quick wins & scalable solutions 18:20 — Guardrails for AI with imperfect data: “cost of being wrong” and risk metrics 21:20 — Human-in-the-loop and deterministic quality checks 23:00 — Human expertise as “moat”; why experience in using AI matters 26:30 — Workflow integration: Why domain experts’ role changes in the agentic AI era 28:45 — Abhishek’s definition of agentic AI & digital FTEs (AI agents) 31:00 — Product innovation: 5-stage process to commercialise ideas (ideation, validation, experimentation, incubation, commercialisation) 33:00 — Measuring ROI from innovation: board-level, commercial, and activity metrics 34:30 — Rapid-fire Q&A: Open source, explainability, speed vs. perfection, AI’s biggest risks Thanks for listening to the CAIO Connect Podcast. If you enjoyed this episode, please leave a 5-star review to help spread the word about the show. Be sure to subscribe so you never miss another insightful conversation. Never forget to subscribe CAIO Connect AI Newsletter for Future AI Updates, News, and Events: https://caioconnect.org/news-and-insights/ Streaming On: Spotify: https://open.spotify.com/show/5AgxOyDPFfoHwEQJVRwBAj Apple: https://podcasts.apple.com/sg/podcast/the-caio-podcast-with-sanjay-puri-chief-ai-officer-podcast/id1749085722 YouTube: https://www.youtube.com/@caioconnectpodcast Join our fastest-growing AI Community: Twitter: https://x.com/caioconnect Instagram: https://www.instagram.com/caioconnect/ LinkedIn: https://www.linkedin.com/company/caioconnect/ Read Our Blogs, News & Updates: https://caioconnect.org/ Join the Conversation: Leave your thoughts and questions in the comments below. We'd love to hear from you! #AIExpert #CAIOConnectPodcast #ChiefAIOfficer