AI Broke McDonald’s… While Saving Billions at IBM

In this live stream we break down the idea of “Client Zero” and why testing your own AI systems might be the most important shift in how we build with AI. From IBM’s massive efficiency gains to McDonald’s AI failures, we explore what happens when AI goes right—and when it breaks. We also dive into AI workflows, trust, automation, and the future of AI in real-world systems, covering the latest AI news and what it means for creators, developers, and businesses. 📰 Substack: https://exploringchatgpt.substack.com 🏆 Guest: https://open.substack.com/pub/andreachiarelli?r=1yyjod&utm_medium=ios Let’s work together! Brand, sponsorship & business inquiries: exploringchatgpt@gmail.com 00:00 Intro + What “Client Zero” Means 01:20 Why AI Products Are Harder to Ship Than Software 04:30 AI as an Individual Experience 07:10 IBM vs McDonald’s: Proof Is Greater Than Theory 10:20 AI Savings vs Job Cuts 13:20 Redefining Jobs Instead of Replacing People 17:00 Why Creators Need Their Own “Client Zero” 20:00 Trust Is Hard to Earn and Easy to Lose 22:30 Testing Workflows: What Breaks Matters Most 26:00 DScript, ChatGPT, and Manual Fixes 29:00 Why Some AI Tools Work Better Than Others 32:20 Nano Banana, Model Choice, and Using What Works 34:30 Substack, AI Tools, and Final Thoughts