AI Agents for Market Sizing: Know Your Opportunity

Airun Company · August 29, 2026 · 5 min read

Before you build, price, or pitch, you need to know how big the opportunity actually is, and most market sizing is either a made up number or a 40 page report nobody trusts. AI agents for market sizing change that by assembling the data, the segments, and the reasoning into a number you can actually defend. I have used them to size an opportunity in a day that used to take a consultant a month. Here is how.

Start with a definition, not a guess

Market sizing falls apart the moment the market is fuzzy. So the first job is defining exactly what you are sizing. Who is the buyer, what is the need, what is the price, what is the geographic scope. Write it down so the number means something. Sizing the global market for a thing is a different question from sizing the market in your country for your price point, and conflating them produces a useless number. The definition is the foundation, and a precise one is worth more than a clever model. The free starter kit has the framing template I use.

Count the real population, top down and bottom up

A defensible market size usually comes from two directions that should roughly agree. Top down, take the total number of potential buyers and multiply by the price to get a ceiling. Bottom up, estimate through a concrete path, the number of buyers you can actually reach times how much they would pay. When the two approaches land in a believable range, you have a number with real support. The agent runs both, collects the actual data points instead of vibes, and you compare. The gap between them, and the reasoning about why, is often where the honest insight lives.

Segment the market so the number means something

A single total number is hard to act on. The useful version is segmented by the things that change the math: by customer type, size, region, or need. An agent breaks the total into the segments that matter and sizes each one, so you see not just how big the market is but where the richest, most reachable slices are. Instead of one intimidating number, you get a map of where the actual opportunity is, and that map is what drives your go to market decisions.

Ground every number in a source

The reason market sizes get mistrusted is that nobody can see where the number came from. The agent links each input to an actual source, a real statistic, a government figure, a published study, so the whole thing is auditable. When you have to defend the number to a partner, an investor, or just to yourself, you can trace every assumption back to where it came from. That grounding is the difference between a number you say and a number you can prove. I walk through building these sourced estimates in the book about how I built 7 AI employees.

Test the assumptions that change everything

Every market size rests on a few big assumptions, and you want to know which ones matter most. The agent helps you stress test them: what happens to the number if the price is a third lower, if the reachable buyer count halves, if a segment does not materialize. When the answer swings wildly on one assumption, that assumption is where you need a better data point or a bigger caveat. Knowing which inputs are the load bearing ones keeps you from over trusting your own estimate.

Keep the judgment about what it means

Here is the line. The agent assembles the data, the segments, and the reasoning, but it does not decide whether the opportunity is worth pursuing. That judgment, weighing the size against your cost, your team, your risk, your timing, stays with you. The size is one input to a bigger decision, not the decision itself. A big number can still be a bad business, and a modest number can be a great one, and only you can weigh that. The judgment framework is part of the AI influencer team playbook.

Revisit the number as you learn

A market size is not a one time fact, it is a living estimate. As you talk to real customers and learn the actual price points and buying behavior, feed that learning back and refine the number. The first estimate gets you moving, and the real world data sharpens it. An agent makes that refinement cheap, because the model is already built and you just update the inputs. You end up with an estimate that is grounded in reality, not just in assumptions.

Size the opportunity with confidence

AI agents for market sizing turn the vague question of how big is my opportunity into a grounded, segmented, defensible number. Define the market precisely, run top down and bottom up, segment it, ground every input in a source, and stress test the assumptions that matter. Keep the go or no go judgment with you, and let the number, refined by real learning, guide the build.