A non-technical founder called me the week after a board meeting, rattled. An investor had asked, almost in passing, "so what is your AI strategy?" and the founder had given a vague answer about "exploring some tools." The investor nodded and moved on, but the founder could feel the ground shift. The question was not hostile. It was the new baseline expectation, and not having a crisp answer read as not being on top of the company's biggest lever and biggest risk.
If you are pre-seed to Series A and you do not yet have a clean answer to that question, here is the short version: your AI strategy is not a list of tools you use. It is a coherent position on where AI creates real value in your business, what it costs, what risk it introduces, and how you will govern it. You can assemble that in an afternoon if you know what belongs in it. This post lays out what belongs in it.
Why the question is suddenly everywhere
Two years ago, "what is your AI strategy" was a question for large enterprises. In 2026 it is a standard part of board conversations and technical diligence, because investors have watched AI move from a novelty to a line item that shows up in both the upside and the risk column. When a board member asks it, they are usually probing three things at once: are you capturing the efficiency AI offers, are you exposed to a risk you have not priced, and do you actually understand your own product well enough to have a view.
The trap is treating it as a technology question when it is a judgment question. Founders who fumble it tend to answer with a shopping list, "we use a coding assistant and an LLM for our chat feature," which tells the board nothing about whether those choices are sound. A good answer sounds like a business position with technology underneath it, not the other way around.
The five things a real answer contains
A coherent AI strategy for an early-stage company fits on one page and covers five areas. You do not need all of them to be mature. You need a defensible view on each.
Where AI creates value in your product
Be specific about the difference between AI that is core to your product and AI that makes your team faster internally. These are different bets with different stakes. If AI is in your product, name the exact job it does for the user and why it beats the non-AI alternative. If it is mostly internal efficiency, say so plainly, that is a perfectly good answer, and it keeps you from overclaiming an "AI moat" you do not have. Investors have gotten sharp about the difference, and claiming a moat you cannot defend is worse than claiming none. I dug into that specific failure mode in what investors really mean when they ask about your AI moat.
What it costs you
If AI is in your product, you need to understand its unit economics, because inference is a variable cost that scales with usage in a way that traditional software does not. A board member who asks about your AI strategy is often really asking whether every new user makes you money or quietly costs you money. If you cannot answer roughly what an AI-powered interaction costs you and how that moves as you grow, you have a hole exactly where the sharpest question will land. This is worth working out before the meeting, not during it, and I walk through the mechanics in the unit economics of an AI feature.
What risk it introduces
This is the half of the answer founders most often skip, and it is the half that separates a credible answer from a naive one. Name the real exposures honestly: data going to third-party model providers, the possibility that a model you depend on gets deprecated or repriced, hallucinated output reaching customers, and the security surface of any AI system that can take actions. You do not need to have solved all of these. You need to show you see them and have a plan proportional to your stage. A founder who says "here are our three real AI risks and here is how we are containing each" sounds far more in control than one who insists there are none.
How you govern it
Governance sounds heavy for a five-person company, but at your stage it means something simple and concrete: do you have a basic policy for what data can go into which tools, who can turn on a new AI vendor, and how you review AI-generated code before it ships. A one-paragraph answer here, "engineers can use approved tools, customer data does not go into unapproved models, all AI-generated code is reviewed like any other code," is enough to show the board that AI is not entering your company through an ungoverned side door.
Where you are deliberately not using it
The strongest AI strategies include a line about restraint. Saying "we are not putting AI into X because the failure cost is too high and the value is marginal" signals judgment better than any list of adoption. It tells the board you are making decisions, not chasing a trend. A founder who can articulate what they chose not to automate is a founder who is actually thinking.
How to assemble your answer without a technical background
If you are non-technical, none of this requires you to become an engineer. It requires you to sit down with whoever owns your technology and turn their working knowledge into a one-page position. Walk through the five areas, write a few honest sentences on each, and pressure-test the two that draw the hardest questions: the value claim and the cost. If you do not have a senior technical person to do this with, that gap is itself part of the answer, and it is a common reason founders bring in a fractional CTO before a raise, to build the position and stand behind it in the room.
The goal is not a polished deck. It is that the next time an investor asks the question in passing, you answer in four crisp sentences: here is where AI matters in our product, here is what it costs us per user, here are the two risks we are managing, and here is what we have deliberately chosen not to do. That answer ends the topic and moves the meeting forward, which is exactly what a good answer to a board question is supposed to do.
FAQ
What should a startup's AI strategy actually include?
A one-page position covering five things: where AI creates real value in your product versus internally, what it costs per interaction, what risks it introduces, how you govern its use, and where you have deliberately chosen not to use it. It is a business judgment document with technology underneath, not a list of tools.
How do I answer "what is our AI strategy" if I am non-technical?
Sit down with whoever owns your technology and turn their knowledge into a few honest sentences on each of the five areas. Focus hardest on the value claim and the cost per interaction, since those draw the sharpest follow-up questions. You do not need to be an engineer, you need a defensible position.
Do investors expect us to have AI in our product?
No. Investors expect a coherent view, not forced adoption. "We use AI internally for efficiency and deliberately keep it out of X" is a strong answer. Claiming an AI moat you cannot defend is worse than honestly saying AI is a supporting tool rather than the core of your business.
What is the most common mistake founders make here?
Answering with a shopping list of tools instead of a business position, and skipping the risk and cost sections entirely. Boards read an all-upside answer as naive. Naming your real exposures and your unit economics is what makes the answer credible.
If your next board meeting or raise is coming and you want a sharp, defensible AI position built with someone who has sat on the other side of that table, you can book a call.