Skip to content

Build a competitive comparison from multiple source documents

Turn a folder of competitor briefs, product pages, and analyst notes into a structured comparison matrix — ready for an exec review or sales enablement.

Stage: Cowork · For: Manager, End user · Level: Intermediate · Time: 20 min

When to use this

You have a competitive review coming up. You've gathered research — product comparison docs, analyst reports, pricing pages, maybe some field notes from sales conversations — but it's scattered across files and formats. Synthesizing it manually into a structured comparison takes hours.

Cowork reads all the source documents in parallel, extracts the relevant dimensions, and produces a comparison matrix or narrative brief that would take a human analyst half a day to build.

What you'll need

  • M365 Copilot license with access to Cowork
  • At least 2-3 source documents per competitor (or a single comprehensive source per competitor)
  • A clear list of the comparison dimensions that matter for your use case

Try it now — the task description

Open Cowork and describe the task:

Build a competitive comparison of [Competitor A], [Competitor B], and [Competitor C]
based on the attached documents.

Compare them across these dimensions:
1. Core product capabilities / what it does
2. Target customer and use case
3. Pricing model (if available)
4. Key strengths
5. Key weaknesses or gaps
6. How they position against us / [your product or service]

Format the output as:
- A comparison table (competitors as columns, dimensions as rows)
- A 2-paragraph narrative summary: where we're stronger and where we're at risk

Audience: [sales team / product team / leadership team].

Why this task format works: Specifying the comparison dimensions before Cowork runs prevents it from choosing dimensions that don't match your decision context. The table + narrative combination serves two use cases — the table for quick reference, the narrative for exec-level communication.

Step by step

  1. Gather your sources. Attach the competitor research docs to Cowork. The more consistent the source quality across competitors, the more balanced the output.
  2. Run the task. Cowork will process all documents in parallel and populate the matrix.
  3. Spot-check the cells. For any cell that seems vague or wrong, ask: 
    What specific evidence from the source documents supports the claim about [Competitor A]'s [dimension]?
  4. Fill gaps. If a dimension is empty for one competitor, ask: 
    Is there any information in the documents about [Competitor B]'s pricing? If not, note it as "not disclosed."
  5. Sharpen the narrative: 
    Rewrite the narrative section to be more direct about our biggest risk and our clearest differentiator.

Screenshots

We deliberately don't ship screenshots that go stale — the Microsoft Copilot UI changes often. Follow the numbered steps above, which we keep current. Maintainers can regenerate fresh captures with the Playwright tool in tooling/screenshots/.

Make it better

  • Win/loss lens: add a section "Based on the positioning differences, what situations favor us and what situations favor each competitor?" — useful for sales coaching.
  • From web research: if you don't have existing docs, ask the Researcher agent to gather competitor intelligence first, then hand the output to Cowork for structuring.
  • Refresh cadence: save the task description as a recipe and re-run it quarterly with updated source documents.
  • Product team variant: add "Feature comparison table: list specific capabilities and mark each competitor as 'Yes / Partial / No' for each." — useful for roadmap prioritization.

Watch out for

  • The matrix is only as balanced as your sources. Thin or dated material on one competitor produces a confidently wrong cell.
  • Cowork won't flag what's missing unless you ask. Empty or vague cells need a “what evidence supports this?” pass.
  • Positioning claims pulled from marketing docs describe how a competitor sells, not how they perform. Keep the two separate.

Where this leads (the ramp)

Running this comparison once is a half-day saved; running it every quarter means rebuilding the same task each time. In Agent Builder you can package the whole competitive-research routine as a Research Librarian agent your team invokes on demand.

Next: Agent Builder: build a Research Librarian agent

Was this page helpful?

Found an issue or want to suggest an improvement? Open a GitHub issue  ·  Star on GitHub