<p>Are you still trying to figure out if agentic AI is hype or reality? </p><p>Alex Taylor, Global Head of Emerging Technology at QBE Ventures, cuts through the noise in this no-nonsense conversation about what&#39;s actually working in insurance AI - and what&#39;s failing spectacularly.</p><p>Discover why agentic AI isn&#39;t just &quot;fancy RPA,&quot; how insurers are running shadow mode tests to prove AI can outperform human underwriters, and why the real barrier isn&#39;t technology, it&#39;s data strategy. Alex shares jaw-dropping examples from software development (27-hour autonomous coding sprints!) and explains how insurers are moving from chatbot failures to genuine operational transformation.</p><p>Key insights: the difference between vibe coding and provable AI, why observability matters more than accuracy, Microsoft-Allstate&#39;s governance playbook, and the one thing every insurance CIO must do in the next 30 days.</p><p>If you&#39;re responsible for AI strategy, digital transformation, or innovation in insurance, this episode delivers the practical framework you&#39;ve been missing. No vendor pitches. Just real talk about implementation, regulation, partnerships, and what separates AI winners from the FOMO-driven crowd.</p><p><br>Timestamps</p><ul><li><p>0:00 - Introduction - Alex Taylor &amp; QBE Ventures</p></li><li><p>1:30 - The shift from &#39;what&#39;s possible&#39; to &#39;what works&#39; in insurance AI</p></li><li><p>2:15 - Why insurers underinvested in technology (and why it made sense)</p></li><li><p>3:45 - The real problems insurers are trying to solve with emerging tech</p></li><li><p>5:00 - Internal pressures: cost, complexity, and competitive speed</p></li><li><p>6:20 - Customer expectations and the value proposition (spoiler: they don&#39;t care about AI)</p></li><li><p>7:30 - What actually changed in the last 12-18 months</p></li><li><p>8:30 - Agentic AI explained: beyond classical generative AI</p></li><li><p>9:45 - The critical difference between agentic AI and RPA</p></li><li><p>11:20 - The operating system experiment: 27 hours of autonomous coding</p></li><li><p>13:00 - Inversion of control: humans as engineering managers</p></li><li><p>14:30 - Build vs buy vs partner: how the calculation has changed</p></li><li><p>16:15 - What the ideal tech stack looks like: people, process, tech, governance</p></li><li><p>17:45 - The regulatory complexity and governance requirements</p></li><li><p>18:30 - Snorkel&#39;s AI leaderboards and model certification</p></li><li><p>19:45 - Case study: What didn&#39;t work (the chatbot mistake 99% made)</p></li><li><p>21:30 - What actually works: agents as employees, not buttons</p></li><li><p>22:15 - Metrics that matter: measuring AI against human baselines</p></li><li><p>23:30 - Shadow mode testing: running parallel systems for 12 months</p></li><li><p>25:00 - Partnership models: how CVCs accelerate experimentation</p></li><li><p>26:30 - QBE&#39;s Lighthouse Program: 3-week proof of value</p></li><li><p>27:45 - Cutting through the hype: what&#39;s real vs. overstated</p></li><li><p>28:45 - The one thing to do in the next 30 days: know where your data is</p></li><li><p>30:00 - Closing thoughts and where to follow Alex&#39;s content</p></li></ul><p><br></p>