01 / START HERE
A useful way to approach it.
A competitor report should inform a specific decision: which audience to serve, what to include in an offer, or which promise needs stronger evidence. Start with three competitors and a narrow question. A list of everything an assistant can find is harder to review than a comparison built around the same criteria.
Use published websites, documentation, and other accessible material. In signed-in o1 chat, current retrieval depends on enabled web search and service availability. You can also supply excerpts yourself. Keep a company's own claims separate from your interpretation, and mark anything you could not verify. The example below uses fictional companies and authored sample evidence.
02 / THE PROCESS
Go from question to next step.
- 01
Define the decision and the comparison set
Write one sentence describing the decision this research should inform. Choose direct alternatives serving a similar audience, then include one indirect alternative if customers might solve the problem differently. Record the region and date. Comparing an enterprise platform with a local service can be useful, but explain that difference before interpreting their offers.
- 02
Build a shared evidence checklist
Choose five or six criteria such as target customer, core job, package boundaries, onboarding, support, and proof offered. Assign a source to each criterion: product pages for published positioning, documentation for described behavior, and case studies for reported customer outcomes. Do not let the most detailed website determine all your categories.
- 03
Collect claims with their context
Capture the URL, page title, relevant wording, and access date. Note whether a statement applies to every package or only a particular tier. If web search is unavailable, paste accessible excerpts and ask the assistant to use only those. A suggested source list is a research plan until somebody opens the sources.
- 04
Separate findings from interpretations
Ask for three distinct fields: what the company publishes, what that may imply for your audience, and what remains unknown. Label marketing language as a published claim. An onboarding video can show a described process; it does not prove every customer follows it or that the process is faster than yours.
- 05
Turn the comparison into one test
Review the weakest evidence before choosing a positioning direction. Identify one question you can investigate through your own customer conversations, a product demonstration, or a small offer test. Keep the public comparison as a dated baseline. It does not reveal competitors' private revenue, roadmap commitments, negotiated prices, or internal strategy.
03 / YOUR STARTING POINT
A prompt you can make your own.
Replace the bracketed fields. Add only information you have permission to share. This is a reusable text prompt; it does not run a task from this page.
Compare {{competitors}} for {{audience}} deciding {{decision}} in {{region}}. Use these public URLs or supplied excerpts: {{sources}}. If enabled web search retrieves additional evidence, link the original pages and record access dates; otherwise mark missing evidence. Compare target customer, core offer, package limits, onboarding, and published proof. For each finding show the source claim, your interpretation, and an unknown. Do not infer that an undocumented feature is absent. Finish with three verification questions and one practical test.04 / SEE THE SHAPE OF IT
Illustrative competitor evidence matrix
Fictional companies and authored source notes; this is a worked example, not live research or an o1 execution.
| Criterion | Fictional evidence | Interpretation / next check |
|---|---|---|
| Customer | Maple: page S1 addresses solo consultants | Check whether teams are also supported |
| Offer | Harbor: page S2 describes a setup service | Different buying choice from self-service |
| Limits | Maple: S3 lists two active projects | Confirm how an archived project is counted |
| Support | Harbor: S4 names email support | Response time is not stated |
The report produces useful questions without pretending that four public pages reveal either company's complete product.
05 / KEEP YOUR JUDGMENT
What to watch before you use it.
Copying positioning as fact
A phrase such as 'easiest platform' describes marketing, not a demonstrated comparison. Record it as the company's claim.
Mixing different packages
A feature on an enterprise page may not apply to the entry offer. Keep tier and eligibility beside the evidence.
Reporting absence too confidently
Use 'not found in the reviewed sources' when documentation is incomplete. Ask a follow-up question before making a stronger conclusion.
06 / THE FINAL PASS
A little review goes a long way.
Your checks stay on this page. They do not approve or run a task.
07 / GOOD QUESTIONS
Before you begin.
How many competitors should I include?
Start with a small set you can inspect carefully. Three direct alternatives often give a manageable first comparison. Expand only when another competitor adds a different audience, offer model, or way of solving the customer's problem.
Can AI discover a competitor's private strategy?
This workflow analyzes accessible evidence. It should not claim access to private plans, customer records, or internal performance. Public messaging may suggest a positioning choice, but the interpretation must remain labeled and open to correction.
What should I do when a source is inaccessible?
Record the gap and use another accessible primary source if one exists. Do not invent the content or present a search snippet as a completed review. You can supply permitted excerpts for a narrower comparison.