Market Sizing Pro.
Hand it a company, get a defensible TAM/SAM/SOM with live sources.
Find Sources subflow: composes search queries, hits Google for market data, dispatches the scrape subflow per source.
Scrape Sources subflow: pulls and cleans the page content for one source URL.
End of run
A full market-sizing deliverable: TAM, SAM, and SOM numbers with the methodology that produced each, the search queries that sourced the data, an executive summary in board-ready prose, and a transparent trail back to the underlying citations. Input is just a company; output is what you’d otherwise commission from a research firm.
$1.68 a report, ~6 minutes per run. 56 runs to date.
Challenges it solves
Real numbers, not vibes. The agent doesn’t ask Claude what the market size is. It scrapes the company’s website to understand the actual product, generates a battery of targeted search queries that surface real industry sources, extracts and simplifies the relevant data, then runs the TAM/SAM/SOM calculations against that. The number ladders up from sourced inputs.
Methodology you can defend. Every output passes through a
Calculate TAM/SAM/SOM step that exposes the formula and the
inputs, then a Create Executive Summary step that puts the
answer in language a board can read. Both come out together, so
the reader can move from headline to math without leaving the
artifact.
Avoid generic search slop. Rather than one Google query,
the agent first analyzes the company, then Generate Search Queries composes multiple angled searches (industry sizing
reports, comparable competitor revenue, government statistics
where relevant), broader coverage, less noise.
Under the hood
Three workflows wire it together. Main.flow orchestrates the
end-to-end run: Start → Collect Company Info → Scrape Website → Analyze Company → Generate Search Queries → Run Find Sources → Simplify → Extract Sources → User Context → Calculate TAM/SAM/SOM → Create Executive Summary → Generate Text → End.
Find Sources.flow composes the queries, hits Google for the
market data, and dispatches the scrape subflow per source.
Scrape Sources.flow is the small utility that pulls and
cleans one URL’s content for the extraction step.