The pool of people who can build a go-to-market function inside an AI company is small, and almost all of them are employed. They are not applying; they are evaluating.

When they take a first conversation with a Series A developer tooling company, they arrive with a diligence list that would not embarrass a junior investor.

This is a change in the market and it is worth taking seriously. The Founders who lose these candidates rarely lose on money. They lose because a question landed and the answer wobbled.

“What Happens to This Product If the Next Model Is Twice as Good?”

This is the question candidates ask most often and Founders answer worst. The instinct is to insist the moat is defensible and move on, which reads as either naivety or evasion.

The credible version accepts the premise and gets specific. A strong answer names the layer where the value sits and explains why model improvement strengthens rather than erodes it – the workflow, the data that accumulates through use, the evaluation infrastructure, the enterprise integration surface, the distribution. It also names what would genuinely break: “if frontier models ship native long-horizon planning with reliable tool use, our orchestration layer becomes a thin wrapper, and here is what we would do in that case.”

Candidates are not looking for certainty. They are testing whether the founder has thought about the failure mode. A Founder who has is someone they can sell alongside.

“Show Me the Runway Maths”

Senior GTM candidates now ask for months of runway, current net burn, and the specific metrics the next round will be raised against. Some ask about the gross margin profile once inference costs are loaded, which in AI companies can be the difference between a software business and a services business wearing software pricing.

“We are well funded” is not an answer. “We have twenty-two months at current burn, the Series B case is built on reaching four hundred thousand of net new ARR per quarter by Q2, and gross margin is sixty-one per cent and improving as we move workloads off the largest model” is an answer. It is also a brief: a good seller immediately understands what they are being hired to produce.

Founders sometimes worry about disclosing this before an offer. Under an NDA at final stage it is normal, and candidates at this level have seen worse numbers than yours. Withholding it signals the numbers are bad.

“Who Owns the Data, and Does That Hold?”

A candidate who has sold enterprise AI before knows that data rights are where deals slow down, and where a company’s long-term advantage either exists or does not. They will ask what the standard contract says about training on customer data, whether the largest customers negotiated that clause away, and whether the roadmap depends on rights the contracts do not actually grant.

The question is not adversarial. It is the candidate working out whether the thing they will be asked to sell is the thing the company has. A Founder who can walk through the standard terms and the two exceptions is demonstrating the operational honesty that makes people join.

“What Are My Shares Actually Worth?”

Equity fluency has risen sharply. Expect questions on the number of options rather than the percentage, total outstanding shares, strike price, last preferred price, liquidation preference structure, and the post-termination exercise window.

A founder who cannot answer these in the room is not disqualified, but the follow-up matters: send the numbers within twenty-four hours. Candidates who cannot value equity assign it a value of zero and negotiate hard on cash — so vagueness here is expensive in a way that is easy to miss.

“Why Did the Last Person in This Seat Leave?”

In a market this small, they will find out anyway. Someone in the candidate’s network has worked with your last VP of Sales or your first AE, and the backchannel conversation happens whether or not you invite it.

The strong move is to get there first and be plain. “We hired a sales leader nine months too early, the motion was not ready, and it did not work. Here is what we changed.” That answer costs nothing and buys a great deal, because it tells the candidate the founder can name a mistake without flinching. The alternative – a vague story about fit – is the version they will later hear contradicted by someone they trust.

The Answer That Loses the Candidate

Across all of these, the losing pattern is consistent: the founder treats diligence as scepticism and starts selling harder. The energy in the room changes, the candidate stops probing, the conversation ends warmly, and the offer is declined a fortnight later for reasons that sound like timing.

Strong candidates are not asking these questions because they are unconvinced. They are asking because they are seriously considering it. Interrogation is interest. Treat it as such.

Preparing the Honest Version

The preparation is not complicated and it takes an afternoon. Write out the five questions above. Write the honest answer to each, including the parts that are uncomfortable. Have the equity numbers and the runway figures in a document you can send the same day. Decide in advance which of the uncomfortable answers you will volunteer rather than wait to be asked.

Then say the honest version out loud to someone who will tell you if it sounds rehearsed.

The founders who win these hires are not the ones with the cleanest story. They are the ones whose story survives contact with a well-informed sceptic — which, increasingly, is every candidate worth hiring. We prepare founders for exactly these conversations before the first candidate call: how we work.