Tariq Choudry of Amazon Web Services talks about why AI pilots still fail, cyber risk, decisions over dashboards, & why AI will replace heroics, not humans. IN THIS EPISODE WE DISCUSS: [04.13] An introduction to Tariq, his background, and role at AWS. "I spend my time thinking about how we move from software that explains problems to software that actually solves them at scale." [06.18] Why AI will replace heroics, not humans. "Supply chains are held together by caffeine, guilt, that one person that hasn't had a vacation since 2019. There are a lot of late nights and Slack war rooms, and there are groups of people that have the entire network in their hands. That's extremely fragile – and not scalable." [10.10] Why so many AI pilots still fail, what's going wrong with both technology and people, and the big problem with incentive and blame culture. "Pilots don't fail because the underlying model is bad. They fail because the organizations are very good at protecting how decisions are currently made. Companies are saying they want AI – but only if nothing important changes." "If all you're doing is trying to determine what failed, why, and who's to blame, you've missed the point." [15.30] How businesses can incorporate new capabilities and integrate them into their existing systems and workflows, and use agentic AI to surface the need for critical decisions earlier when there's more time and optionality. "Time is the one commodity you can't earn back… Use the agent to surface those weak signals earlier – that's when you still have options." [21.17] From dashboards and Excel to tribal knowledge in our workflows, how AI is exposing organizational debt, and what that means for teams. "You spend your time fighting the fires, and less time designing the new systems to prevent them." [26.49] What does all of this means for planners? "The best planners won't get replaced – they should be promoted!" [30.43] Why cyber risk is now a supply chain problem, and how AI can helps teams navigate it. "Your weakest supplier is your weakest point in your firewall." [33.39] Why people want AI but don't trust it, and why trust is built from predictability. "When humans make mistakes, over time we call that judgement. It comes from experience – that's a judgement call. But when AI makes that mistake, it's scandalous." "Trust isn't perfection, it's predictability." [38.37] Tariq's advice for how businesses can build trust in AI, prove predictability, and scale with confidence. RESOURCES AND LINKS MENTIONED: Head over to Amazon Web Service's website now to find out more and discover how they could help you too. You can also connect with AWS and keep up to date with the latest over on LinkedIn, Facebook, YouTube, Instagram or X (Twitter), or you can connect with Tariq on LinkedIn. If you enjoyed this episode and want to hear more from Amazon Web Services, check out 489: Time To Swap Your Axe For A Chainsaw: The Power of Agentic AI or 519: Overcoming The Perfect Storm: Moving Beyond Basic Automation To Realize AI's Full Potential. Check out our other podcasts HERE.