When an AI tool gives a bad answer, the instinct is to blame the model or the vendor. More often the culprit is sitting in the CRM, in the form of duplicate accounts, missing fields, conflicting definitions and data spread across systems that were never designed to talk to each other. This panel gets into the unglamorous work that makes AI useful, from cleaning and connecting siloed data to deciding which sources to trust. Expect real examples from organisations of very different shapes and sizes, and frank views on what to fix first when you can't fix everything at once. Key takeaways: - The most common data problems that undermine AI tools, and how to diagnose them - How to prioritise data fixes that will have the biggest impact on AI results - Practical ways to connect siloed data for better planning, segmentation and prioritisation