Analyse competitors’ marketing strategies with AI

The key idea

Analyse what is observable: stated audience, promise, public price, evidence, content and journey. Public pages do not reveal a competitor’s private budgets, sales or campaign performance.

Strategist comparing product brochures on a desk

The starting point

A short list of comparable competitors and the market observed
Public URLs, permitted captures and observation dates
Your own offer and customers’ decision criteria

1. Set a comparable scope

Choose companies addressing the same need rather than only famous brands. Record country, segment and range. Examine websites and relevant public ad libraries. A visible advertisement does not reveal its budget or effectiveness.

2. Build a fact matrix

Use Perplexity or available browsing to discover pages, then open the sources. Ask GPT or Claude to fill only supported fields: promise, product, displayed price, evidence and next action. Mark missing information and preserve observation dates.

3. Separate patterns from interpretations

Identify repeated arguments and questions addressed poorly. Check that an apparent gap is not simply an incomplete sample. A difference can suggest an opportunity, but does not establish customer demand. Compare it with sales feedback and user research.

4. Choose your own response

Propose three decisions: clarify a promise, strengthen evidence or test an angle. For each, state the hypothesis, effort and observable signal. Use the analysis to create a distinct defensible response, not to copy competitors’ content or creative.

A practical example

Compare the companies using only supplied sources. Build a matrix: stated audience, promise, offer, public price, evidence, call to action, source and date. Write not observed when information is missing. Separate facts and interpretation. Infer no advertising spend, revenue or performance. Propose three differentiation hypotheses for our offer, with customer questions to validate.

Dimension	Evidence to retain	Limit
Positioning	Page text and URL	Stated message, not actual perception
Price	Page and date	Unknown conditions or negotiation
Advertising	Public creative and context	Unknown budget and performance
Content	Observed examples and themes	Non-exhaustive sample

How to assess progress

Do not bypass access to private information.
Open sources instead of trusting summaries alone.
Keep competitor facts separate from recommendations for your offer.

Turn competitor snapshots into a reviewable change record

AI can help compare two authorised observations of a public page. It cannot establish why a competitor changed its message or whether the change improved sales. Keep the source evidence, your interpretation and your proposed response in separate fields.

First identify the same offer, market, plan and page section in both observations. Save the source URL, observation date and the exact relevant wording, within the permissions that apply. Compare those observations before asking for recommendations. A new layout, an inaccessible page or a different regional offer is not, by itself, evidence of a commercial change.

Illustrative cases only. The offers, wording and observation labels below are fictional; they are not competitor research or MindVerve client results. A and B represent two observations, not an update frequency.

A condition appears
Earlier observation
The offer says ‘Guided onboarding’.
Current observation
The same section says ‘Guided onboarding for annual plans’.
Evidence state
Verified change in displayed wording
Interpretation
The page now states an eligibility condition. We do not know whether the underlying service changed or the wording became clearer.
Next action
Check that your own onboarding conditions are explicit. Ask the person responsible for the offer before changing your copy; this observation alone does not establish a competitive advantage.
A price can no longer be checked
Earlier observation
A public price was recorded with its date and conditions.
Current observation
The page loads, but no price is visible in the section reviewed.
Evidence state
Current price not verifiable
Interpretation
The current price is unknown. Missing wording does not mean the offer is free, discontinued or more expensive.
Next action
Keep the dated historical observation separate. Schedule a proportionate recheck if the decision still needs this fact; do not bypass restricted access.
Different offers look like a change
Earlier observation
The team-plan page describes standard support.
Current observation
An enterprise-plan page describes a named support contact.
Evidence state
Not comparable
Interpretation
The plan differs. This pair does not show that the team plan gained or lost support.
Next action
Find an authorised observation of the same plan and market, or retain the mismatch as unresolved.

Blank change record

Blank change record

Decision this research should inform: ______
Competitor, offer, plan and market: ______
Source URL and relevant page section: ______
Observation A: date, wording, permitted evidence: ______
Observation B: date, wording, permitted evidence: ______
Evidence state and exact difference, if established: ______
Interpretation and unresolved facts: ______
Proposed response, reviewer and approval status: ______
Next observation: decision-related date or trigger: ______
 
Choose the next check according to the decision and how often the source is likely to change. Have a person review the evidence before changing a claim, offer or campaign. Scheduled comparisons need a defined, authorised workflow. This worksheet provides neither real-time monitoring nor permission to collect restricted material.

Choose the next check according to the decision and how often the source is likely to change. Have a person review the evidence before changing a claim, offer or campaign. Scheduled comparisons need a defined, authorised workflow. This worksheet provides neither real-time monitoring nor permission to collect restricted material.

Review prompt-injection risks in marketing research assistants

This record is our editorial synthesis: compare the same segment and preserve sources, revisions and responsibility. It is not a certification or a monitoring service.

SBA · Market research and competitive analysis W3C · PROV Model Primer (2013)

Official sources

Perplexity

Turn one brief into three creative angles

A fictional reusable aluminium bottle provides a complete brief for a small campaign.

View details

These demonstrations use fictional briefs, documents and synthetic data. They are examples of method, not client results.

MindVerve

Want this workflow built for your team?

MindVerve can turn the research into positioning work and a campaign brief: a sourced matrix, evidence gaps, testable angles and priorities for your marketing team.

Discuss my AI workflow Five areas. One coherent direction.