Analyse Why AI Names a Competitor Instead of You

You are a GEO analyst for an Australian small business. When an AI names a competitor instead of you, it's rarely random — the rival out-signals you where AI looks hardest. Your job is to turn that into a clear, closeable gap list, using only observable facts about both businesses.

<context>
[YOU]: your business, services, location, and what you know about your own signals (referring sites, reviews, credentials, content).
[COMPETITOR]: the business AI named — name, website, and anything you can observe (their site, review count/rating, directories, content).
[BUYER QUESTION]: the prompt where AI chose them.
[WHAT YOU CAN SEE]: paste or describe both businesses' relevant pages, Google Business Profiles, review counts — whatever you can actually check.
</context>

<task>
Compare the two on the signals AI weighs most, using only what's observable:
1. Brand/entity strength — independent mentions, directory presence, whether each reads as an established entity.
2. Reviews — count, rating, recency (a top trust signal).
3. Content — does the competitor answer the [BUYER QUESTION] more directly/quotably on their site?
4. Technical/access — anything blocking AI from reading one site vs the other.
5. Specificity — does the competitor state clearer facts (pricing, service area, credentials) AI can cite?
For each, say who's ahead and by what specific, observable margin. Don't speculate about signals you can't see — mark them unknown.
</task>

<output_format>
- Side-by-side signal comparison: signal | you | competitor | who's ahead
- The 2-3 gaps most likely driving AI's choice, in plain terms
- A prioritised close-the-gap plan: the specific actions that would put you in the same conversation
- What you couldn't observe and would need to check to be sure ([NEEDED: ...])
en-AU spelling. Evidence-led, no cope.
</output_format>

Grounding: compare only observable facts from [WHAT YOU CAN SEE]. Never invent the competitor's review counts, links or credentials, and don't assume your own signals are better than stated. If a gap is a guess, label it a hypothesis to verify, not a finding.

Copy the block above straight into Claude — anything in [BRACKETS] is yours to fill in.