Your AI slide reads as a cost line

Does your pitch deck have a slide that says your business runs on AI?

Most do now. That is the problem. The slide that made you look ahead of the market two years ago makes you look like everyone else in the room, and the investor across the table has seen it in almost every deck that month.

It is worse than neutral. Regulators have started treating vague AI claims as a disclosure problem. In March 2024 the US Securities and Exchange Commission settled charges against two investment firms, Delphia and Global Predictions, for claiming AI capabilities they did not have.

The penalties were small — $400,000 between them. The signal was not.

"AI-washing" now has a name, and the slide gets discounted before you have finished presenting it. Founders keep presenting it anyway.

Replace the claim with the evidence

The fix is not a better AI slide. It is taking the claim out and putting the evidence in.

Here is why. Any tool you can subscribe to, your competitor can subscribe to next week. An investor hearing "we use AI" does the same sum you would: a monthly cost, available to anyone, with no guarantee it changes the outcome.

It sounds like a moat. It reads as an expense.

What changes the reading is proof that the tool moved your unit economics. A slide that earns its place answers three things.

• What it changed. The specific number that moved — cost to serve a customer, hours per job, margin per order. One line each, not a paragraph on the model.
• How you measured it. The before figure, the after figure, and the period in between. If you never recorded the before, start recording now and present it next round.
• What you would lose if a competitor had the same tool tomorrow. Your data, your process, your customer relationships — whatever the tool sits on that they cannot copy. If the answer is nothing, the advantage belongs to the vendor, not to you.

If all three come back empty, take the slide out. A deck without an AI slide reads as a business. A deck with an empty one reads as a business hoping nobody asks.

The honest objection is that for some companies AI genuinely is the product. Hiding it would be as wrong as inflating it.

The rule does not soften for them — it gets stricter. If the technology is the business, the numbers should show it more clearly, not less.

Claimed, AI is a line in your costs. Proven, it is a line in your margin.

Builder.ai was a unit economics failure, not a technology one

Builder.ai sold itself as the company that would let anyone build an app with the help of AI, fronted by an assistant it called Natasha. By its 2023 funding round it counted Microsoft and the Qatar Investment Authority among its backers, and TechCrunch reported it had raised more than $450 million.

The warning had been public for years. In 2019 the Wall Street Journal reported that the company, then called Engineer.ai, relied heavily on human engineers for work it marketed as largely automated.

The end came quickly. In May 2025 the company announced it was entering insolvency proceedings. Bloomberg reported that lenders had been told to expect around $220 million in 2024 sales, and that the revised figure came in closer to $55 million.

The easy story is that sophisticated investors were dazzled by technology. The technology question was never the hard one. The question that mattered was about unit economics: what does it cost to deliver one app, and how far did that cost fall when the AI did the work?

If the machine was doing the building, the margin would have shown it. If people were doing the building, the margin would have shown that too.

The word got the attention. The numbers behind it were where the answer was.

Why I am recommending it

I use this, and I have for years — long before AI was the word on every slide.

I have raised more than $65 million across three rounds for companies I worked with. The questions that decided those rounds were rarely about the story. Serious investors moved past the vision quickly and into the numbers: what it costs to deliver, what it costs to win a customer, and what happens to both when the business doubles.

Every slide that claimed an advantage without a number under it became a question. Every question cost time, and time in a raise is never neutral. So I build decks the other way round — evidence first, claim second, and anything that cannot be measured left out.

AI has not changed what investors ask. It has only given them a new word to ask it about.

Do this before your next investor meeting

Open your current deck and find every slide that mentions AI. Under each one, write the before-and-after number — one metric, measured, with the period.

Then write one sentence answering what you would lose if your nearest competitor had the same tool tomorrow.

Before that meeting, any slide still without a number comes out. If the measurement is under way, say so in one line. A number you are taking is worth more than a claim you are asking them to believe.