Ask most AI tools about a company and you get the same four facts. What it does, how big, when it raised, recent headlines. If you're deciding whether to build on their API, what matters is how often it breaks. No profile contains that.

Strawberry researches it the way you would with an afternoon free. Reading the docs, working through tools you already have access to, checking sources you wouldn't have thought of. What comes back is a brief you could defend in a meeting, with a source next to every claim.

Start with the question, not the company

The same company supports very different investigations. Sizing up a potential vendor means looking hard at reliability, support, pricing history, and who else already depends on them. Understanding a competitor who has just entered your market is closer to product detail, hiring, and where their attention is going. Reading up on a company before a first conversation may need nothing more than a current, accurate picture of what they actually sell.

So it helps to say what you are deciding and how much the answer is worth. A bounded question should come back quickly with something useful. A consequential one deserves a second pass into the places where the evidence is still thin.

Ask your Strawberry companion to research the company around that question, and it will work through the sources with you rather than handing back a profile.

Want to try it?

Ask your Strawberry companion: “Research this company around the question I need answered. Use relevant context and current sources, then give me a concise brief with the important findings, changes, tensions, gaps, and source links.”

Skill

Research a company

Research this company around the question or decision I need help with.

Read past the marketing pages

A company's marketing pages tell you what it wants understood, which is rarely the thing you came to find out. The more useful material usually sits further in, a changelog that shows how fast the product really moves, documentation that reveals what the integration will actually require, a status history, a pricing page read against the version archived last year, a support forum where the same complaint keeps reappearing.

Most of that is reachable but tedious. Some of it sits behind a login you already have: an analyst database your team subscribes to, a procurement portal, a shared drive holding last year's evaluation. Strawberry can work through both, following a promising link several levels down, searching inside a site rather than around it, and using the sources you are authorized to use, then bring what it finds together with the public web. You don't have to work out in advance which sources will matter and paste them in.

Keep the gaps and contradictions visible

A brief earns its length by being clear about what it knows. It should be obvious which claims are the company's own, which are corroborated somewhere independent, which are inference, and where the evidence simply isn't there yet.

  • The clearest answer to the question you started with.
  • The company context needed to understand that answer.
  • Important changes, patterns, or tensions.
  • Source links and dates beside the findings that matter.
  • Contradictions, gaps, and useful questions to investigate next.

Contradictions are worth more than they first appear. When a company's job postings and its published roadmap disagree about where the product is going, that gap is often the most useful line in the brief, and it only shows up if the two sources are read against each other rather than summarized separately.

Know when it's really a sales or investment question

A general brief stops at a clear, source-linked view of the company. Some decisions carry their own discipline behind them, running a deal, testing a fund's thesis, and those are better served by the workflow built for them than by one brief stretched to cover everything.

Finish the research before you design the deck

Research and design are separate jobs. When they run together the research tends to lose, because slides that look finished are easy to mistake for questions that are settled. Once the company view is accepted, Strawberry can build the presentation from it without reopening the investigation.

Skill

Turn an accepted brief into a beautiful deck

Turn this accepted company brief into a polished, source-grounded presentation.

If your team researches companies the same way repeatedly, the accepted sources, questions, depth, and brief format can be saved as a shared skill. A recurring update can become a Routine once the sources, the changes worth tracking, the cadence, and the review point are clear.

How to map a market with AIMove from one company to the categories, relationships, and patterns shaping a broader landscape.
Official Strawberry skill
Copy skill file Download skill file
name
research-a-company
description
Research one company into a concise, source-linked brief focused on the question or decision in front of the user. Use for general company understanding, partner or vendor research, strategic research, or a current company view that is not primarily a sales account brief, marketing competitor analysis, or investment screen.

Research a Company

Help the user understand one company well enough to answer a real question or make a better decision. Adapt the depth and output to that question instead of filling a standard company profile.

1. Identify the company and the useful answer

Confirm which company the user means, resolve ambiguous names or domains, and understand what the research should help them decide. They may need a quick current view, an explanation of the business, an assessment of a potential partner or vendor, an understanding of recent changes, or a deeper answer about the product, team, strategy, customers, ownership, or market position.

Use the time available and likely consequence to set the depth. Give a useful first answer when the question is bounded; propose a deeper pass when missing evidence could materially change the view.

2. Start with relevant context and authorized sources

Use approved tabs, files, prior research, notes, connected apps, and internal documents when they help answer the question. Do not ask the user to reconstruct context that is already available.

Research current primary and credible secondary sources. Strawberry may be able to reach useful sources beyond ordinary public search by working with sites and tools the user is already authorized to access. Follow promising links, search within sources, and investigate past the first generic result when greater depth could change the answer. Never bypass access controls or treat technical availability as permission to use sensitive information.

Choose sources for the question rather than visiting the same checklist every time. Useful evidence may come from company and product pages, documentation, filings, announcements, leadership material, hiring, customer or partner references, reviews, reporting, databases, industry sources, and approved internal context.

3. Reconcile the company view

Confirm identities before combining records. Keep material sources and dates close to the claims they support. Distinguish the company's own claims, observed facts, third-party evidence, internal context, calculations, and inference.

Look for changes, tensions, and missing evidence that matter to the user's question. Do not force a recent signal, silently resolve contradictions, or present a public proxy as proof. Stop when the question is answered or more searching is unlikely to change the decision.

4. Deliver the brief the decision needs

Lead with the answer or most useful current view. Depending on the question, include:

  • what the company does and the context needed to understand it;
  • the findings most likely to change the decision;
  • important recent changes, supported patterns, or tensions;
  • relevant people, products, customers, ownership, partnerships, or market context;
  • contradictions, uncertainty, and evidence still missing; and
  • useful next questions or the next research step.

Return a concise, source-linked brief in the format the user needs: chat, document, table, spreadsheet, or another agreed destination. Prefer a few decision-relevant findings over a long, generic profile.

5. Hand off to the specialist owner

Use strawberry/sales/research-an-account when the company is being researched as an account and relationship for a sales decision. Use strawberry/venture-capital/research-an-investment-opportunity when thesis fit and the investment screen are the primary result. Use strawberry/marketing/run-competitor-analysis for a marketing competitor comparison, or strawberry/research-analysis/map-a-market when the question concerns a broader landscape rather than one company.

Research and presentation production remain separate jobs. When the accepted findings should become a presentation, use strawberry/research-analysis/create-beautiful-slide-deck without reopening the research by default.

After feedback, offer to preserve trusted sources, depth, questions, fields, and output shape as a custom or team skill. Suggest a Routine only when the company, source set, change definition, cadence, audience, destination, and review point are stable.