AI Is Cheating Until It Reaches The Cap Table

Everyone Is Asking the AI Question

I am asked about AI regularly by investors, partners, and business leaders. And by regularly, I mean every single day for as long as my patience will allow.

They should ask. I do.

Sometimes the person asking knows exactly what he or she is asking. That is a very rare event.

Sometimes, somebody received a directive that every management meeting now requires the AI question.

“What is your AI strategy?”

Check.

Next slide.

Many people on the other side of the glass are talking about AI, instead of learning it.

I say that with love. Kind of.

Checking for the Noun

“What is your AI strategy?” sounds smart. In the hands of people who are not empowered to make strategy and will not be held accountable for the result, it can also be useless and destructive.

Remember the questions that got us here: Which client problem are we solving? What efficiency are we creating? How do we measure whether it worked? What old work goes away? Where does human judgment stay? What happens when the model gets it wrong? Is the improvement sustainable?

And now, more than ever:

What happens to the person driving the AI question when it works, or when it doesn’t? Who owns it?

Accountability has to go full circle. It cannot stop with the employee, partner, or leader who was told to “go make AI happen.”

That does not mean every experiment must work. It means the owner chooses a real problem, runs a disciplined test, measures honestly, learns quickly, and kills a bad idea before it becomes permanent. The person applying the pressure also owns the quality of the question.

Especially when that pressure began with a memo, grammar-checked by Grammarly, written by ChatGPT, and based on market intelligence Claude assembled from articles that may also have been written with AI

By the time it reaches the operator, half the chain may be artificial, but somehow all the accountability is human.

When the implementation fails, how quickly is the owner expected to learn, change course, or kill it?

“Looks good” does not cut it anymore. AI can make every memo polished, every slide look intelligent, and every half-formed idea look like someone spent six months developing it.

AI can make the presentation better without making the idea any less stupid.

The Money and the AI Consultants Have Arrived

Bain surveyed⁠ private investors representing $3.2 trillion in assets under management. They reported that a majority of their portfolio companies were testing or developing generative AI, and nearly 20 percent had operationalized use cases and were seeing concrete results.

Bain called that impressive for such a nascent technology. Fair enough. It is.

Its larger conclusion is the right one. The firms making progress are applying AI to strategic priorities, requiring quantified benefits, monitoring relevant metrics, and treating AI as a tool in service of strategy, not as the strategy itself.

Those are not noun checks. Those are real business initiatives.

The money has arrived.

The board questions have arrived.

The AI consultants have definitely arrived.

Over the years, consultants have sold us servers, digitization, the cloud, OCR, machine learning, automation, and drag-and-drop customization. Now, overnight, many of them market themselves as AI consultants.

Some are excellent. Some simply changed the title slide.

Different deck. Same invoice.

The question is not whether investors, boards, CEOs, or AI consultants should push companies to move faster on AI. We should.

The question is whether that pressure is attached to a real business result and whether everyone applying it is willing to share responsibility for what happens next.

AI strategy is not a mandatory slide, a license count, or an executive saying “AI” three times during an earnings call and hoping the valuation fairy appears.

It is a series of accountable business decisions: a real problem, a named owner, a measurable result, clear boundaries around human judgment, a plan for what happens when the model is wrong, and a commitment to remove the old work when the new approach succeeds.

The market may pay for the promise.

The people paid to perform still have to produce the proof, preferably in the P&L, not the PowerPoint.

So please keep asking the AI question, but make sure the work it triggers is tied to a real business problem or opportunity, has a named owner, and produces a measurable result.

Otherwise, you are not doing diligence.

You are simply an empty suit checking a box to confirm that the noun “AI” appeared somewhere in the deck.

To learn more about how this philosophy has driven success, visit my website.