Oct. 21, 2026

248: Drew Mehta: "Why Most AI Projects Deliver Faster Garbage Instead Of Business Value"

248: Drew Mehta: "Why Most AI Projects Deliver Faster Garbage Instead Of Business Value"
I Have Some Questions...
248: Drew Mehta: "Why Most AI Projects Deliver Faster Garbage Instead Of Business Value"

Drew Mehta explains why so many “AI strategies” fail in practice, and why owner-operators need end-to-end workflow decomposition, not deck-first consulting. We dig into the difference between consultants and builders, the “garbage equation,” and why AI should be used to fix weeks-long process bottlenecks, not just minutes of work.

👤 About the Guest

Drew Mehta is an AI and product builder with 20+ years of experience creating data, analytics, and AI products for Major League Baseball, KKR, and PepsiCo. He founded ThinkFraction to help small owner-operated businesses ship AI-enabled improvements by decomposing workflows, measuring impact, and pairing advisory with implementation and education.

🧭 Conversation Highlights

  • How consulting incentives (deck deliverables, time and materials) can create a strategy-to-execution gap in enterprises
  • Why AI “faster garbage” happens when leaders optimize the wrong thing, leaving queue time and handoffs untouched
  • The emerging FDE (forward deployed engineer) mindset, and why AI makes that hybrid role more valuable. It’s basically what the technical product manager role has been at many firms for many years. New name, a few distinctions playing the same game.
  • Leadership and learning: cheap signals get cheaper with AI, but trust and development still require real human feedback and mentorship


💡 Key Takeaways

  • If AI only speeds up the minutes, but the process still takes weeks, you did not fix the problem.
  • Consulting and building need different incentives: decks don’t equal shipped ROI.
  • Decompose workflows first, then decide what to change across people, process, and technology.
  • AI can force better self-diagnosis, but leadership still depends on high-quality human signals and feedback.


❓ Questions That Mattered

  • Why do large companies so often split strategy from implementation, and how does that mismatch harm outcomes?
  • What does it mean to avoid “the garbage equation” when adopting AI?
  • How should owners and leaders think about governance vs guardrails when using AI with sensitive data?
  • As AI cheapens communication signals, what becomes the risk to trust and leadership development?


🗣️ Notable Quotes

  • “Broken process plus AI equals faster garbage.”
  • “If AI cuts the work from 60 minutes to 30 minutes, but it still takes 22 days, that’s the problem.” (Credit to CEO of Varick Agents)
  • “The deliverable for a consultant is a deck. The builders ship and demonstrate quantifiable impact.”
  • “A review is qualitative, not just quantitative.”

🔗 Links & Resources