AI stalls when boards treat it as a technology problem. Why governance, shared data, and leadership accountability are what determine AI success.

In December 2025, I audited my own time. What I found was sobering.

Sixty-eight percent of my calendar had been consumed by management overhead: follow-ups, status checks, chasing data across siloed systems, stitching together fragmented views of reality just to make basic decisions. I was doing what many startup CEOs do without admitting it: living in a state of strategic whiplash.

So, I built something. Over five days, I constructed Triton, a multi-agent system connecting our CRM, product management, communications, and financial data into a single operating layer. For the first time in my career, I had real-time visibility into the heartbeat of the business without waiting for someone to assemble a report. Board reporting that used to consume weeks started taking 11 minutes. The time I got back allowed me to build again, including Mighty Humans, a free AI literacy program for displaced workers in partnership with AWS and Anthropic.

I am not sharing this as a productivity story. I am sharing it because of what it revealed about leadership, and what too many boards are making impossible for CEOs to do.

AI does not respect org charts

AI does not move through the organizational chart; it moves across it, cutting across functions, data, decisions, all the places where ownership gets murky. That is precisely where most boards are not looking.

The old structure, where CTO, CMO, CFO, and CRO each run their own kingdom, made sense when functions were largely self-contained. Today, those borders are liabilities. The companies pulling ahead are not the ones with the best tools. They are the ones with the most clearly defined operating models: shared records of decisions, shared metrics everyone trusts, and horizontal data access. This is not an IT problem, but a governance problem, and it sits squarely at the board level.

The board's job is not to recommend tools, but to define how the organization operates around them. That means asking different questions: not which AI vendor did we select, but do we have a single source of truth for revenue? Do our functional leaders share metrics? Do we know what happens to jobs when a process gets automated, and have we communicated that clearly?

Most boards are not asking those questions yet. And because they are not, every function is buying its own tool, building its own silo, and calling it transformation.

The adoption gap is a people problem

Some of the largest technology companies in the world are sitting at around 20% AI adoption internally.¹ Companies that build AI products cannot get their own people to use them. That is not a failure of technology. It is a human one.

I have seen this pattern up close. CEOs declare AI a priority, and then hand over the mandate to functional leaders and move on. Most of those leaders are not equipped for it. They have not been given clarity on what transformation means for their teams, what happens to the roles that get automated, or what the path forward looks like for the people doing those jobs right now.

So, people do what humans always do when they face an existential threat. They slow down, hedge, say the right things in all-hands meetings, and then keep doing what they were doing.

Words do not matter right now. Actions do. Unilever is training tens of thousands of employees on AI, not just deploying tools at the top and hoping adoption trickles down.² That is a board-level commitment. Someone in that organization decided the human condition was not a problem to be managed, but an investment to be made. Most boards are still waiting for their functional leaders to figure it out. That wait is expensive.

Your data problem is a leadership problem

When I connected Triton to our systems of record, I did not start with a data cleaning project. I started with a trust problem. Which numbers did we believe? Which metrics did everyone agree on? Where were we carrying three different versions of the truth, one in sales, one in finance, one in the board pack, without saying it out loud?

That is the conversation most companies avoid. And when you put AI into an environment like that, it does not resolve the disagreement but accelerates it. You get faster, more confident versions of three different wrong answers.

Imagine: if everyone in your organization was working from the same numbers, what would that change?

AI learns from documentation. Better-structured companies compound their advantage faster. The fix is not a new data platform, but a board that insists on shared records, definitions, and cross-functional accountability for the numbers. That has always been a governance question. AI just makes the consequences of getting it wrong arrive faster.

Brand is your next moat. Make it machine-readable.

Governance does not stop at internal operations. Brand has become an operating question, and most boards are still treating it like a marketing one.

The first engagement a potential customer has with your company may never involve a human being. AI agents are already reviewing vendors, comparing feature sets, narrowing options, and shaping shortlists before a human decision-maker ever enters the room. Brand needs to work at two levels now: memorable for humans, and legible for machines.

The companies that treat brand as a board agenda item instead of a marketing one will have a disproportionate advantage in the next five years. One company I work with made exactly that decision. Revenue outperformed expectations by roughly 19%, share of voice climbed by around 26%, and they recorded their highest profits in 18 months. That is not a coincidence. That is what happens when brand strategy gets the same boardroom seriousness as financial strategy.

The human cost is also a business cost

Every time I sit in a boardroom or around a kitchen table right now, I hear the same thing: fear about what AI is going to do to jobs. That fear is real and legitimate.

If you build or deploy technology that displaces roles, you carry equal responsibility for what comes next for the people in those roles. Not as a PR exercise. As a genuine strategic and moral commitment.

That is why I started Mighty Humans. It came directly out of the capacity I freed up. It also came from a conviction: if you are going to participate in building the next era of work, you have to participate in the transition, too. Boards that treat purpose as separate from performance will be corrected by the market faster than they expect.

This is a human moment

Everything I have talked about can be turned into a framework: a maturity model, a readiness assessment, or a risk register. All of that has its place.

But what I found when I audited my own time was not a productivity hack. It was a reorientation of where my time goes, what I owe to the people around me, and what I want to build with whatever capacity I reclaim.

The board's job right now is to give their organizations that same permission. To define how the business operates in this new era. To hold leaders accountable for making the hard calls, not just the easy ones. And to ensure that no one, from the executive team to the frontline, is left to figure this out alone.

This is not a moment for governance frameworks. It is a moment for turning up, being present, and making the human calls that no AI can make for you. The ones about people, about obligation, about what kind of organization you want to be on the other side of this.

No model can make those calls. That is still the job.

References:

• Deloitte. (2026). The state of AI in the enterprise. Deloitte US. https://www.deloitte.com/us/en/what-we-do/capabilities/applied-artificial-intelligence/content/state-of-ai-in-the-enterprise.html
• Unilever. (2025, December 1). How digital transformation drives our operational excellence. https://www.unilever.com/news/news-search/2025/how-unilevers-digital-transformation-is-driving-operational-excellence/