How to Create a Feedback Loop for AI Innovations

Prompting isn’t the hard part anymore. Trusting the output is. You finally get a model to reason step-by-step… And then? You're staring at a polished paragraph, wondering:    > “Is this actually right?”    > “Could this go to leadership?”    > “Can I trust this across markets or functions?” It looks confident. It sounds strategic. But you know better than to mistake that for true intelligence. 𝗛𝗲𝗿𝗲’𝘀 𝘁𝗵𝗲 𝗿𝗶𝘀𝗸: Most teams are experimenting with AI. But few are auditing it. They’re pushing outputs into decks, workflows, and decisions— With zero QA and no accountability layer 𝗛𝗲𝗿𝗲’𝘀 𝘄𝗵𝗮𝘁 𝗜 𝘁𝗲𝗹𝗹 𝗽𝗲𝗼𝗽𝗹𝗲: Don’t just validate the answers. Validate the reasoning. And that means building a lightweight, repeatable system that fits real-world workflows. 𝗨𝘀𝗲 𝘁𝗵𝗲 𝗥.𝗜.𝗩. 𝗟𝗼𝗼𝗽: 𝗥𝗲𝘃𝗶𝗲𝘄 – What’s missing, vague, or risky? 𝗜𝘁𝗲𝗿𝗮𝘁𝗲 – Adjust one thing (tone, data, structure). 𝗩𝗮𝗹𝗶𝗱𝗮𝘁𝗲 – Rerun and compare — does this version hit the mark? Run it 2–3 times. The best version usually shows up in round two or three, not round one.  𝗥𝘂𝗻 𝗮 60-𝗦𝗲𝗰𝗼𝗻𝗱 𝗢𝘂𝘁𝗽𝘂𝘁 𝗤𝗔 𝗕𝗲𝗳𝗼𝗿𝗲 𝗬𝗼𝘂 𝗛𝗶𝘁 𝗦𝗲𝗻𝗱: • Is the logic sound? • Are key facts verifiable? • Is the tone aligned with the audience and region? • Could this go public without risk? 𝗜𝗳 𝘆𝗼𝘂 𝗰𝗮𝗻’𝘁 𝘀𝗮𝘆 𝘆𝗲𝘀 𝘁𝗼 𝗮𝗹𝗹 𝗳𝗼𝘂𝗿, 𝗶𝘁’𝘀 𝗻𝗼𝘁 𝗿𝗲𝗮𝗱𝘆. 𝗟𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 𝗜𝗻𝘀𝗶𝗴𝗵𝘁: Prompts are just the beginning. But 𝗽𝗿𝗼𝗺𝗽𝘁 𝗮𝘂𝗱𝗶𝘁𝗶𝗻𝗴 is what separates smart teams from strategic ones. You don’t need AI that moves fast. You need AI that moves smart. 𝗛𝗼𝘄 𝗮𝗿𝗲 𝘆𝗼𝘂 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝘁𝗿𝘂𝘀𝘁 𝗶𝗻 𝘆𝗼𝘂𝗿 𝗔𝗜 𝗼𝘂𝘁𝗽𝘂𝘁𝘀? 𝗙𝗼𝗹𝗹𝗼𝘄 𝗺𝗲 for weekly playbooks on leading AI-powered teams. 𝗦𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 to my newsletter for systems you can apply Monday morning, not someday.

My AI lesson of the week: The tech isn't the hard part…it's the people! During my prior work at the Institute for Healthcare Improvement (IHI), we talked a lot about how any technology, whether a new drug or a new vaccine or a new information tool, would face challenges with how to integrate into the complex human systems that alway at play in healthcare. As I get deeper and deeper into AI, I am not surprised to see that those same challenges exist with this cadre of technology as well. It’s not the tech that limits us; the real complexity lies in driving adoption across diverse teams, workflows, and mindsets. And it’s not just implementation alone that will get to real ROI from AI—it’s the changes that will occur to our workflows that will generate the value. That’s why we are thinking differently about how to approach change management. We’re approaching the workflow integration with the same discipline and structure as any core system build. Our framework is designed to reduce friction, build momentum, and align people with outcomes from day one. Here’s the 5-point plan for how we're making that happen with health systems today: 🔹 AI Champion Program: We designate and train department-level champions who lead adoption efforts within their teams. These individuals become trusted internal experts, reducing dependency on central support and accelerating change. 🔹 An AI Academy: We produce concise, role-specific, training modules to deliver just-in-time knowledge to help all users get the most out of the gen AI tools that their systems are provisioning. 5-10 min modules ensures relevance and reduces training fatigue.  🔹 Staged Rollout: We don’t go live everywhere at once. Instead, we're beginning with an initial few locations/teams, refine based on feedback, and expand with proof points in hand. This staged approach minimizes risk and maximizes learning. 🔹 Feedback Loops: Change is not a one-way push. Host regular forums to capture insights from frontline users, close gaps, and refine processes continuously. Listening and modifying is part of the deployment strategy. 🔹 Visible Metrics: Transparent team or dept-based dashboards track progress and highlight wins. When staff can see measurable improvement—and their role in driving it—engagement improves dramatically. This isn’t workflow mapping. This is operational transformation—designed for scale, grounded in human behavior, and built to last. Technology will continue to evolve. But real leverage comes from aligning your people behind the change. We think that’s where competitive advantage is created—and sustained. #ExecutiveLeadership #ChangeManagement #DigitalTransformation #StrategyExecution #HealthTech #OperationalExcellence #ScalableChange

AI workflows will never be your moat. A thoughtful data loop might be. I've noticed a consistent blind spot where founders obsess over model architecture but neglect the systems that make those models continuously better. Here's the hard truth: your competitors can license the same foundation models, fine-tune with similar techniques, and replicate most of your AI capabilities within months. What they can't easily replicate? Your data. The most successful AI companies build what I call "inevitable data loops" - systems where: → Data collection is woven seamlessly into the user experience → Users willingly contribute data because they receive immediate value → Each interaction improves the product for all users → These improvements drive both retention and word-of-mouth growth → This creates a flywheel effect that accelerates over time. Consider the difference: WEAK DATA STRATEGY: "We'll ask users to rate outputs and occasionally send feedback" STRONG DATA STRATEGY: "Our workflow captures implicit signals about which outputs users actually implement, creating a continuous training signal without additional user effort" From an investor and operators perspective, you should be thinking: → How will you capture proprietary data that others can't access? → Why will users give you this data without friction? → How exactly does this data feed back into product improvement? → What prevents competitors from building similar data advantages? The best founders have clear, specific answers to these questions - not vague gestures toward "collecting feedback." Remember that most AI experiences degrade over time without fresh data. Your initial performance means little compared to your rate of improvement. The question isn't just what your product can do today, but how it gets better tomorrow — systematically and effortlessly. Building valuable AI products is fundamentally about designing elegant data capture mechanisms; not just clever implementations. #startups #founders #growth #ai