You can build a useful market landscape in an afternoon if you define the boundary before you collect the companies.

Draw the edge of the map before searching for names.

A landscape of “AI companies” is a universe, not a research task.

A landscape of European tools that automate back-office documents for insurers, sold to mid-market operations teams, has a boundary someone can test.

Strawberry can help decompose the question into search tracks and source types. Specify the entities that count, the adjacent categories that do not, and the geography. Also specify whether the goal is investment, product strategy, partnership scouting, or account targeting.

Build the company list from evidence, not search-result rank.

Search results favour companies with good SEO, not necessarily companies with traction or relevance. Check a company’s own site, product pages, pricing, customer material, company profile, and current announcements before you decide where it belongs.

An agent can gather those sources in parallel and record why a candidate passed the boundary. Preserve the URL and date checked so the map remains inspectable when someone asks why a name is present.

Group companies by the choice a buyer would actually make.

A segment should help someone compare alternatives, not merely sort companies alphabetically.

You might group by buyer, workflow, deployment model, data source, or position in the value chain, depending on the question behind the map.

Strawberry can suggest patterns in the evidence and build a draft matrix, but you should name the categories deliberately. If a company fits two segments, capture that ambiguity rather than forcing a false precision.

Finish with what the map changes and what it cannot answer.

A market landscape is useful when it informs a next move.

That might be which segment to validate, who to interview, where the competitor set is crowded, or which partners to research. It is less useful when it ends as a beautiful static slide with no decision attached.

Save the source list and segmentation logic as a landscape skill if the market needs monitoring. A periodic routine can detect new entrants, category moves, funding, pricing shifts, or launches, while keeping unverified signals out of the main map.

Frequently asked questions