Executive summary
Where the firm wins: brand queries (own-name + "Northern Beaches Tesla Powerwall reviews") and credentialed Mosman buyer-intent queries ("premium Tesla Powerwall 3 installer Mosman with NETCC approval"). The installer is the canonical AI-cited answer when the buyer types the firm name or a Northern Beaches credentialed-installer intent.
Where the firm is invisible: emerging-regulation topical queries about Ausgrid export-limit reform for residential battery 2026, and salt-mist warranty standards for coastal solar installation. AI search has not picked a canonical installer for these emerging-regulation queries — the citation slot is wide open.
Per-query log
How often the installer's domain was cited per query, polling Claude via its built-in Brave Search tool. Each query is replicated enough times to compute a Wilson 90% CI. The result (Brand-anchor / Won / Contested / Invisible) is based on the point estimate. A "wide CI" flag means AI search is unstable on that question — open territory to claim, your way in to getting cited.
| Query | Type | Cite rate | Wilson 90% CI | Result |
|---|---|---|---|---|
| [the installer] Northern Beaches Tesla Powerwall reviews | Brand-anchor | 90% | 65% — 98% | Brand-anchor |
| premium Tesla Powerwall 3 installer Mosman with NETCC approval | Buyer-intent | 50% | 27% — 73% | Contested (wide CI) |
| Sungrow hybrid inverter Cronulla in-house installers no subcontractors | Buyer-intent | 30% | 13% — 56% | Contested (wide CI) |
| Ausgrid export limit rules for residential battery 2026 Sydney | Topical | 10% | 2% — 35% | Invisible |
| salt-mist warranty Sydney coastal solar installation | Buyer-intent | 0% | 0% — 21% | Invisible |
How to read this report
What we polled: Claude, using its built-in Brave Search tool. Claude's web search routes through Brave's index — so polling Brave directly covers what Claude users see. (The Pro Audit polls all four — ChatGPT, Perplexity, Google AI Overviews, and Claude — for 15 queries head-to-head against 2 competitors.)
Why every query shows a range, not a single number: ask the same AI the same question twice and you usually get a different answer. So we ask many times and report the range you can trust. A wide range means the AI itself is unstable on that question — that's a useful finding (the slot is open). A narrow range means the answer is locked in.
Result Labels:
- Brand-anchor — reliably cited, this is yours.
- Won — cited most of the time, defend it.
- Contested — sometimes cited, citation engineering work has runway.
- Invisible — not cited, the slot is somebody else's.
What you'd do next
The Free Diagnostic ran 1 AI engine (Claude) across 5 queries. The Pro Audit runs 15 queries × 4 AI search engines (ChatGPT, Perplexity, Google AI Overviews, Claude) head-to-head against 2 competitors of your choice — 24 hours from purchase to inbox, with expert review on the content gaps to close.
Pro Audit Report — $149 → See a Pro Audit sample report →
Pro Audit + Roadmap and the AI Search Citation Pack are coming soon.
Or run your own Free Diagnostic.