Ibm Analytics Insights Podcasts

Send us Fan MailIn case you missed it.....RAG didn't just survive the LLM boom — it became the backbone of how enterprises put AI to work. In this replay, we sit down with Douwe Kiela, the AI researcher who led the team that introduced Retrieval-Augmented Generation in 2020 and went on to co-found Contextual AI. Douwe unpacks whether RAG is here to stay, why most enterprise AI dies in the gap between a great demo and production, how to tame hallucinations, and what actually separates a real "agent" from the buzzword. A clear-eyed conversation on building AI that's grounded, useful, and safe to deploy.00:46 Introducing Douwe Kiela 01:37 RAG - Here to Stay or Go? 06:59 LLMs with Context 08:20 Making AI Successful 10:34 Why Contextual AI? 17:18 LLM versus SLMs 20:28 Speed over Perfection 22:07 Hallucinations 26:02 Making AI Easy to Consume 28:50 Defining an Agent 32:53 Reaching Contextual AI 33:14 The Contrarian View 34:37 The Risks of AI 36:53 For FunLinkedIn: linkedin.com/in/douwekiela Webs