The $40 AI Test Every Board Should Take

by Steve Schmith

Two instances of the same structural failure, eight years apart. What building a personal operating system for $40 taught me about AI governance, pilot purgatory, and the one question every board should be asking.

The number that haunted me

In 2018, I was part of research that produced a number that got considerable attention: 2.4 million.

That was the projected number of US manufacturing jobs that would go unfilled between 2018 and 2028. Not because the jobs didn’t exist, but because people with the skills to fill them didn’t. The annual Manufacturing Institute manufacturing skills gap study, produced in partnership with the National Association of Manufacturers and drawing on related manufacturing competitiveness research done in partnership with the World Economic Forum put an economic price tag on it: $454 billion of manufacturing GDP at risk in 2028 alone. $2.5 trillion over the decade.

CEOs were nearly unanimous in conversations informing those competitiveness-related research initiatives. The primary driver behind the skills gap wasn't retirement or a weak economy. It was this: "shifting skill sets due to the introduction of new advanced technology and automation." The same research made a second point just as clearly: talent-driven innovation was the number one driver of manufacturing competitiveness, and access to that talent was critical. The gap wasn't just a workforce problem. It was a competitiveness problem.

I remember often in those research cycles making personal what we were learning from the conversations with CEOs, policy makers, educators, labor leaders and scientists.

For context, I started my education in nursing school. A skilled trade by any honest definition. So, I knew what retraining cost: not just money, but years. When I tried to imagine being 35 years old working in professional services, I couldn't get past how long a path back to nursing would take. Even just thinking about it felt daunting.

Eight years later, I closed my own AI skills gap in two weeks and less than $100.  That’s less than what most organizations spend on a single AI training module. The speed was astonishing.  What I felt was childish.

2026: Same structure, every sector

The technology has evolved at a blistering pace. AI has quickly replaced robotics as the primary disruptor. The dynamic has not.

In February 2026, Slalom-sponsored research published in Harvard Business Review: 93% of leaders and employees say underdeveloped AI skills and a lack of access to training are limiting their organization’s progress. At the same time, 68% of those same people say they’re keeping pace with AI just fine.

The confidence gap is not a perception problem. It is a measurement problem. Organizations confuse access with capability. They have AI tools, training curricula, and pilots. What many don’t have — validated by several guests we’ve had on Inside CVC — are senior leaders and decision makers who have built anything with the technology they’re governing.

The pattern is identical to 2018. A new wave of technology arrives faster than the workforce can absorb it. The gap widens. The economic cost compounds. In 2018, that story belonged to manufacturing. In 2026, it belongs to every sector simultaneously.

Pilot purgatory is a capital problem

If you sit on a board or run a portfolio, you’ve seen this: an AI initiative gets funded, a pilot launches, twelve months later it’s still a pilot. The slide deck gets updated. The business case gets refreshed. The capability gap it was designed to close is still open.

The culprit isn’t the technology or the vendors selling it. It’s an organizational machine that has redefined “AI capability” to mean “AI access” — distributing tools, commissioning training curricula, calling it transformation. None of that produces leaders who can translate between a business problem and a technical solution. It produces leaders who are informed about AI without being capable in it.

Alec Coughlin, an enterprise AI strategist who has advised some of the world’s largest organizations on scaling real AI capability, calls this “pilot purgatory.” On a recent episode of Inside CVC, his advice wasn’t about governance frameworks. It was more direct: get your hands dirty. Move from conceptual understanding to firsthand experience — not as a sponsor of AI initiatives, but as someone who has actually built something.

Joanna Massey, a board governance strategist and Inside CVC guest, sees the skills gap creating a team culture problem that boards aren't yet measuring. When some people on a team are building with AI and others aren't, a new kind of cultural divide surfaces — questions about the authenticity of work, fairness of contribution, and who deserves credit for what. That tension doesn't produce the productive friction innovation requires. It produces the kind that shuts it down, with direct consequences for the board's ability to drive shareholder value.

Harvard Business Review’s March–April 2026 cover story, “Why Great Innovations Fail to Scale,” identifies the mechanism. Innovations fail not because ideas are wrong, but because organizations can’t bridge the gap between domain expertise and technical execution. Leaders who haven’t built anything with AI can fund it and govern it. They can’t lead it.

Pilot purgatory is not a technology failure. It is a leadership capability failure.

Vibe coding. It felt like Legos.

I had a specific problem. My professional life runs across multiple lanes at once: building a global brand, launching a podcast platform, tracking industry signals, managing relationships, developing content, staying on top of a pipeline. I’d tried every obvious combination — Outlook, Teams, OneNote, Slack. Nothing talked to each other. I was the integration layer. What I wanted was a single system built around how I actually work.

So I built it. Two $20 monthly subscriptions and less than $100 in usage fees. Claude for thinking through architecture and refining prompts. Replit’s AI agent to build. Two weeks.

What happened surprised me emotionally before it surprised me technically.

I found myself going to my laptop at all hours. I’d spend 20 minutes getting something out of my head, hit enter, and walk away. By morning, the thing I’d imagined was functional. Or deep, hours-long sessions building and rapidly iterating one specific idea or function.

The feeling was familiar. It took me a few sessions to place it.

It was exactly what it felt like to build with Legos as a kid. It felt childish. It felt amazing.

Excited. Curious. Creative. Imagining. The sense that whatever I could picture, I could build, and the building itself would show me things I hadn’t pictured yet. And virtually no barriers between the idea and the attempt. Like putting on a new pair of sneakers and wanting to run. You know the feeling.

Those aren’t just fond childhood memories. Curiosity, creativity, and the capacity to imagine possibilities are among the most well-documented characteristics of effective leaders. The leadership development industry spends billions trying to cultivate them. No module has ever produced them the way a real build does.

Vibe coding did.

The question every board should be asking

The 2018 research got it right: skills gaps close through doing, not knowing. The AI skills gap will close the same way.

When I vibe coded, I was forced to hold two worlds at once: the business problem I understood deeply, and the technical execution I was learning in real time. I had to translate between them. That translation capacity is exactly what the HBR research identifies as the difference between innovation that scales and innovation that stalls — and exactly what most senior leaders don’t have, because no training program produces it.

In a board discussion, the leader who has built something asks different questions: not whether the technology works, but whether the business problem is defined precisely enough to produce a useful solution, where the incentive seams are, where legacy complexity will absorb the initiative. These are judgment questions. They develop only through building under real constraints.

For boards and investors, this is a diligence signal. Founders and operators who have built with AI make categorically better decisions about where to apply it and how to scale it. The $40/month build is a more reliable indicator of AI judgment than the $2 million pilot.

So here is the question every board should be able to answer: Among the people setting our AI strategy, who has actually built something with the technology they are governing? Not purchased it. Not piloted it. Built it.

If no one in the room can answer that, the strategy belongs to people governing what they have never experienced. That gap is the root cause of every pilot that never became a product.

It starts closing the same way mine did — with a real problem, a small budget, and a laptop at 11pm. It costs $40 a month. And it feels exactly like Legos.

Steve Schmith is Chief Marketing & Ecosystem Officer at u-path and co-host of Inside CVC, a podcast engaging senior corporate leaders globally on venture, innovation, and strategic capital.

References

Deloitte and The Manufacturing Institute, 2018 Skills Gap and Future of Work Study — in partnership with NAM; draws on WEF Manufacturing for Growth research. https://themanufacturinginstitute.org/research/2018-deloitte-and-the-manufacturing-institute-skills-gap-and-future-of-work-study/ and https://www3.weforum.org/docs/WEF_ManufacturingForGrowth_ExecutiveSummary_2013.pdf.

Slalom, “Close Your Workforce’s AI Skills Gap by Designing an Adaptive Organization” — Harvard Business Review, February 2026. https://hbr.org/sponsored/2026/02/close-your-workforces-ai-skills-gap-by-designing-an-adaptive-organization.

Hill, Tedards & Wild, “Why Great Innovations Fail to Scale” — Harvard Business Review, March–April 2026. https://hbr.org/2026/03/why-great-innovations-fail-to-scale.

Alec Coughlin on Inside CVC: “AI Operating Systems, Digital Labor & the End of Pilot Purgatory.” https://www.buzzsprout.com/2466712/episodes/17455521.

Joanna Massey on Inside CVC: “Rethinking Governance in a Digital Age.” https://www.buzzsprout.com/2466712/episodes/18909675