Essays · The AI-Native Founder
The 20 Questions VCs Ask AI Founders, and How to Answer Each One
Free guide 1 of the AI-Native Founder library. Adapted from the playbook I use with founders before they raise: hundreds of founders, over $500 million raised.
The deck gets you the meeting. The questions decide the cheque. Most founders prepare the first forty minutes and improvise the last twenty, and the last twenty are where an investor finds out whether you understand your own business.
For an AI founder the questions have a sharper edge, because the investor starts from a suspicion: that what you built can be copied, that it costs more to run than it earns, and that the “AI” in the pitch is doing marketing work. Every answer below is written to meet that suspicion with evidence.
How to use this guide. For each question, rate your confidence from 1 to 10, write down the evidence you would need to settle it, and practise the answer out loud. Spend your time on the questions that matter most and score lowest. A good answer has four moves: answer directly, name the concern behind the question, back it with something checkable, and connect it to your plan.
1 · The team
1. “Why are you the right team to build this?” What they test: founder-market fit. Have you lived the problem, and can this team build the product? Answer with: the years you spent inside the problem, what each founder built before, the gap you know you have and who fills it. For an AI company, name who owns the model work and who owns the customer.
2. “What happens if you are not the right CEO as the company grows?” What they test: ego and self-awareness. Answer with: that you are attached to the company’s success, not the title; what you are doing to grow as a leader (a coach, an advisor, a specific skill); and that you would work with the board if the company needed someone else.
3. “Why do you want to spend the next ten years on this?” What they test: staying power. Answer with: the personal reason you care, and a picture of where the company is in year five that is bigger than the first product.
2 · Traction
4. “What traction do you have, and why not more?” What they test: progress, and whether you can judge it honestly. Answer with: the numbers, then the real constraint. “Enterprise pilots take a quarter, and I am the only seller” is a better answer than optimism. Then say what the money changes.
5. “Is your early traction representative of the wider market?” What they test: whether your first customers are friendly outliers. Answer with: what your early customers have in common with the next hundred, and at least one customer who found you without knowing you. For AI products, separate pilots from paying contracts.
3 · The market
6. “How big is this market, and why now?” What they test: insight into timing, not a big number from a research report. Answer with: a bottom-up size (customers you can name, times what they pay today) and the specific change that makes now possible: a model capability, a cost that fell, a rule that changed.
7. “How will this market change, and what could kill your opportunity?” What they test: strategic thinking. Answer with: the two or three threats you watch, starting with the one they are thinking of: the model provider shipping your product. Say what you would do in each case.
8. “How do you stay on top of this market?” What they test: whether you will see a shift before your competitors. Answer with: your actual sources: the customer calls you run every week, the people you talk to, what you track. Concrete beats “I read a lot”.
4 · The product
9. “Is this a painkiller or a vitamin?” What they test: whether the customer must have it. Answer with: what the customer does today without you, what it costs them, and what happened when a pilot customer lost access.
10. “What does the product actually do, and why can’t existing tools do it?” What they test: whether you can explain it in two sentences, and whether you know the alternatives. Answer with: the job it finishes, for whom, and the specific thing the current tools cannot do. Bring one workflow, before and after.
11. “Is this a feature a bigger company will add?” What they test: the most common fear about AI startups. Answer with: what you own that a platform does not: a workflow, proprietary data the customer lets you use, distribution, or an outcome you are accountable for. If the honest answer is “our prompts”, work on that before the meeting.
12. “Do you actually need AI, or is it a buzzword?” What they test: substance. Answer with: what the model does that rules could not, how you measure it (a fixed set of test cases and its pass rate over time), and where a person stays in the loop.
5 · The money
13. “How do you make money, and what are your unit economics?” What they test: whether you have thought about the business, not just the product. Answer with: the price, what it is tied to, and your cost per successful task, including the inference, the retries and the human review. AI products carry compute costs on every use, so investors expect lower margins than classic software and want to see the trend.
14. “How will you use this money, and what will it prove?” What they test: whether the round buys the next round. Answer with: two or three milestones that remove the biggest risks, the date you expect to hit each, and the budget behind it.
15. “Can this scale profitably?” What they test: whether growth makes the economics better or worse. Answer with: what happens to cost per task as volume grows (routing routine calls to smaller models, caching, pricing), and how much of your revenue is software versus services, with that share falling.
6 · Go-to-market
16. “What is your go-to-market, and why does it fit this customer?” What they test: whether you know how this customer buys. Answer with: who signs, who uses, how long the sale takes, and the channel that has already worked at least twice.
17. “How do you scale beyond founder-led sales?” What they test: whether the early wins were you, not the product. Answer with: what you have written down about how you sell (the questions, the objections, the price), and the first hire or the first repeatable channel that takes over part of it.
18. “Will customers come on their own?” What they test: whether you know that a good product does not sell itself in B2B. Answer with: the plan beyond product-led growth: the partners, the content, the outbound motion, and the one you are testing now with its result so far.
7 · Defensibility
19. “What stops someone else, especially a big company, from copying this?” What they test: durability. Answer with: the moat you have at this stage and the one you are building next. A data moat counts only if your contracts give you the data rights and results improve as the data grows.
20. “Why hasn’t this been solved before, and why will you win?” What they test: market history, and your unfair advantage. Answer with: why earlier attempts failed (often the cost or capability that only changed recently), and what you know or have that they did not.
Before the meeting
Pick the five questions you would least like to be asked. Answer each one out loud, record it, and listen back. Then rehearse against a model: give it your deck and your numbers, tell it to play a partner who assumes your company can be copied in a weekend, and ask it to put the five questions to you one at a time. The answers you give with a story instead of a number are your homework.
The goal is not a perfect script for twenty questions. It is to show an investor that you understand your business well enough to answer anything they ask, including the questions that are not on this list.
📬 This is free guide 1 of the AI-Native Founder library. One practical piece for early-stage founders building with AI arrives every Friday. Subscribe free.
Mo Ahmed is the author of The Inside-Out Entrepreneur and writes The AI-Native Founder from inside the build.