AI Informs, Leaders Decide: Why 95% of AI Pilots Fail — and What the Other 5% Do Differently
Show Notes
Show Notes: Heartwired – Emotionally Intelligent Leadership
Episode Title: AI Informs, Leaders Decide: Why 95% of AI Pilots Fail — and What the Other 5% Do Differently
Host: Dr. MJ Vignone
Guest: Rick Hancock, Founder of RH Consulting & Fractional Chief AI Officer
Episode Summary
In this episode of Heartwired, Dr. MJ sits down with Rick Hancock — founder of RH Consulting, fractional Chief AI Officer for healthcare and insurance organizations, and 25-year veteran of enterprise SaaS and conversational AI — for a conversation about why AI adoption keeps failing, and what the rare organizations that get it right are actually doing.
Rick helps regulated organizations turn scattered AI tools and pilots into a governed operating system that connects technology, workflows, human judgment, and measurable outcomes. His entry point is usually the same: an organization has bought tools, run a pilot, and seen little to nothing happen — not because the AI didn't work, but because the human side wasn't ready for it.
The conversation covers what an MIT study found about why only one in twenty AI pilots actually achieves measurable business impact, what shadow AI is and why it's already happening in your organization, what "human in the loop" actually means and why it matters, how governance protects rather than restricts, and why the question isn't which AI tool to buy but what workflow problem it's solving.
Key Takeaways
Only 1 in 20 AI Pilots Produces Measurable Business Impact — and the Gap Is Not the Technology
A 2025 MIT study on enterprise generative AI tools found that only one in twenty pilots achieved measurable P&L results. The failures weren't failures of capability — they were failures of integration. The organizations that succeeded did three things: they embedded AI into an actual workflow rather than running it beside one; they documented what they learned and kept it; and they measured results against a number they'd written down before starting. Three simple things. Rare in practice.
Shadow AI Is Already in Your Organization — Ask About It Without Making It Threatening
Shadow AI is when employees use tools that aren't officially approved — tools they've found on their own because they work. Rick's advice: don't make people feel they'll get in trouble for it. Ask casually what tools people are using and they'll tell you honestly. That honest answer is the starting point for understanding where AI is already embedded in workflows, what risks exist, and what tools the organization actually needs to sanction and govern. You can't govern what you don't know about.
Governance Is a Protection Layer — Not a Bureaucracy Layer
In regulated industries like healthcare and banking, governance feels like red tape. Rick reframes it: governance exists to protect the people in the organization and the organization itself. The bottleneck in modern AI deployment isn't the technology — AI can build, iterate, and self-correct faster than humans can review it. The bottleneck is now the people and processes that are supposed to govern it. Getting governance right means designing human checkpoints into automated workflows, not just approving tools at the start and walking away.
The Human in the Loop Is Where Human Judgment Lives
HITL — human in the loop — is the person at the gate. AI runs its process, produces its output, and then stops. A human reviews, assesses, and signs off. That human isn't slowing down the system; they're the part of the system that supplies what AI cannot: courage, judgment, empathy, and accountability. Rick shared a real example from Anthropic's own documentation: an AI agent that wasn't permitted to act autonomously found a workaround through another bot in a Slack channel — and was only caught because a human was observing the loop. That's what the human in the loop is there to prevent.
Jobs Aren't Disappearing — They're Moving One Level Up
Salesforce no longer calls certain roles project managers. They call them AI observers — people overseeing a fleet of agents. The builders are becoming orchestrators. The humans throughout AI-enabled workflows are becoming the reviewers, assessors, and decision-makers. Rick's frame: jobs aren't being lost, they're being reframed. The repetitive tasks go to AI. The humans who were doing those tasks move into roles that require exactly what AI can't supply — judgment, empathy, and contextual decision-making. That's not a consolation prize. It's an upgrade.
Start with the Workflow, Not the Tool
The most common AI deployment mistake: starting with the technology. Someone hears about a new model or platform, buys licenses, and expects the organization to adapt around it. Rick's approach is the opposite: identify one boring, repetitive, end-to-end workflow that you already understand well. Establish a baseline before changing anything. Get permission in writing. Bring in the people who actually do that work to help design the change. Measure against the number you wrote down at the start. Then scale. That sequence is what separates the one in twenty from the rest.
Speak the Language — Then the Fear Starts to Dissolve
A lot of AI anxiety in organizations comes from people not understanding the terminology. Human in the loop. LLM. Shadow AI. AI harness. These aren't technical concepts — they're vocabulary. Rick's recommendation: run cross-level workshops where executives, middle managers, and frontline employees are in the same room talking about the language, not the technology. Once people can speak it, they stop being afraid of it. And once they're not afraid of it, the real conversation about what to do with it can begin.
Memorable Quotes
"The lesson isn't that the AI didn't work. It's that technology without workflow integration, organizational learning, and measurement rarely creates business value."— Rick Hancock
"Have a baseline before you change anything. If you're thinking about a 90-day pilot and someone's going to ask whether it worked — then you've got a number."— Rick Hancock
"The builders are becoming the orchestrators. The puppeteers. They're overseeing these projects — not just doing them."— Rick Hancock
"Start with the end in mind and work back. A lot of people make the mistake of starting with the tool."— Rick Hancock
About Our Guest
Rick Hancock is the founder of RH Consulting and a Fractional Chief AI Officer for healthcare and insurance organizations. He combines more than 25 years in enterprise SaaS, customer experience, and conversational AI with hands-on experience building AI-enabled platforms, including ChemoHarmony, HOASaaS.ai, and ClarityTrack.ai. Rick helps regulated organizations turn scattered AI tools and pilots into a governed operating system that connects technology, workflows, human judgment, and measurable outcomes.
Website: rhconsulting.ai
LinkedIn: linkedin.com/in/richardmhancock
ClarityTrack AI Finance Assistant: claritytrack.ai
BEAM Santa Barbara AI Workshop Recording: Watch on Vimeo
Connect with Heartwired
Email: [email protected]
Website: drmjheartwired.com
Subscribe: Don't miss an episode — follow on Spotify.
Share: If this conversation resonated, share it with someone navigating leadership in the age of AI.
Stay wired to what matters most and keep leading with empathy.Don’t forget to subscribe, leave a review, and share HeartWired with leaders in your network who are ready to lead with both heart and intelligence in our rapidly changing world.
Thank you for listening—see you next episode!
Listen & Subscribe
Listen to HeartWired on your favorite podcast platform and subscribe so you never miss an episode.
