MIT research recently landed with a gut punch: 95% of AI pilots fail.
The stat has been bouncing around my LinkedIn feed as well as HR and tech circles.
But here’s the thing - if your org is still “piloting” AI in 2025, you’re already behind.
By the time a pilot wraps up, the stack has already shifted. What you scoped six months ago is obsolete today.
In a space moving this fast, pilots aren’t just inefficient. They’re paralyzing.
💪🏽 So, what do the 5% get right?
They aren’t running tests in sandboxes. They’re building in the flow of work.
Across the Huertanomics network, the companies actually making progress share a simple playbook:
Philosophy first. A clear north star for how AI ties to business goals. Everyone understands the “why” before touching the “how.”
Open stack. No gated pilot tools. Shared infrastructure that anyone in the org can build on.
Time to build. Not side projects or nights/weekends. Explicit time and space to experiment.
Real use cases. Skip generic vendor demos. Start with 1–2 pain points and solve them directly.
🎙️ Don’t miss tomorrow’s AMA with Andrew Golden
Andrew is opening up his playbook for successfully diving into his ‘AI for HR’ journey.
Bring your questions. Learn how he overcame challenges many HR leaders are facing. And see what this means for the future of HR leadership.
Okay… but how do you actually start?
Here are 5 simple tactics to move from pilot mode → vibe coding:
⭐ Get a demo. Reach out to someone already vibe coding (including me) and ask to see their system live.
🤝 Partner with tech. Grab an engineer or your CTO for an overview of your current stack. Knowing what’s already in place unlocks new build options.
📊 Leverage your data. Start with your pulse or survey data. Pick one painful people process and redesign it with AI.
👥 Find a buddy. Accountability groups and HR communities are full of peers in the same boat. Pairing up accelerates learning.
⚡ Experiment today. Sign up for VO or Lovable and drop a prompt to build a system. The point isn’t perfection - it’s momentum.
This doesn’t mean AI adoption is a free-for-all. Governance still matters. Strategy still matters.
But the orgs making real progress aren’t stuck in pilot mode. They’re creating systems that scale, iterate, and evolve.
We call this shift vibe coding: scrappy, AI-powered systems designed by the people who actually use them.
It’s less “proof of concept.” More “proof in production.”
If your org is still piloting, you risk falling behind.
If you empower people to build, test, and own systems directly—results can be immediate and transformative.
👉 We’ll unpack this further in tomorrow’s How I Built It AMA. See how HR leaders are cutting through the noise and shipping AI systems in real time.
Until then: stop piloting. Start building!
- Stephen