What ChatGPT, Google AI and Perplexity tell buyers and vendors about your firm — measured across 75 real buyer and seller questions, 12 August 2026.
Every quote below is verbatim from a logged, dated AI conversation run under a documented protocol. This is the material your next vendor reads before they ever ring an agency.
New Zealand's portals have already moved: Trade Me explicitly opens its listings to AI crawlers, and OneRoof publishes sitemaps AI can read. Your inventory reaches AI-assisted buyers whether you act or not — what's contested is whose name, whose facts and whose advice frame the answer. Right now that space is being filled by portals, aggregators, calculator sites… and whichever competitor moves first.
Scores are the share of high-intent AI answer moments in which your firm appears, weighted by how close each question sits to a transaction. Every underlying conversation is logged and reviewable.
Even the market leader holds barely a third of the answer space — most of it belongs to portals, government guides and aggregators. That is the opportunity: this ground is winnable, and nobody has claimed it yet.
You publish no rates, so AI builds every fee answer from calculator and referral sites — and quoted three different structures for you in three runs. The most commercially sensitive fact about your firm is currently authored by strangers.
Branch count, rental book, regional expansion — all stale or missing in AI answers. Your growth story is invisible at exactly the moment a vendor asks “how big are they?”
Your firewall turns away AI retrieval systems even where your crawler policy welcomes them — so live AI lookups of your brand pages fail, and the AI falls back to whatever third parties say.
Questions about methods of sale, market timing, due diligence — dozens of measured answer slots — are framed by portals, banks and government guides. One competitor's local market report was the only agency content to frame an entire AI answer in our baseline.
When vendors ask “which agency should I use?”, AI increasingly routes them to referral services and review platforms — intermediaries who charge you for the lead your own visibility should have earned.
Targets are not promises — they're what the assets of today's best-performing firms demonstrably achieve, applied to your gaps. Reachability and accuracy are engineering: fast and largely in your control. Visibility compounds: citations earn citations, and early movers set the baseline everyone else is compared against.
What we won't promise. Nobody controls a language model's output, and anyone guaranteeing you a #1 AI ranking is selling snake oil. What we control is everything the AI reads: your reachability, your facts, your evidence, your content in the measured question slots — and we re-measure monthly against the same 75-question instrument, so progress is a number, not a vibe.
Firewall and crawler policy fixed, listing schema and sitemaps rebuilt, llms.txt published — nobody in NZ real estate has one yet. Reachability re-scored to 9+.
Authoritative fee page, firm facts page, office and specialist data structured for retrieval. The error register drains: every stale or invented claim gets a published correction AI can find.
Advice content aimed at the exact question slots this report measured, franchise and office alignment, monthly re-scores with a board-ready trend line — and alerts when an AI starts saying something new about you.
| Engagement | What you get | Investment |
|---|---|---|
| Findings & Recommendations | This report: 75-question audit, four scores, error register, evidence file | NZ$2,000 +GST · one-off |
| Fix Sprint | Phases 1–2 delivered in one engagement, before/after re-score included | Scoped from your findings |
| The Monthly Review | Phase 3: monthly re-scores, ranked report, verbatim-claim exhibits, what to fix next | NZ$750 +GST / month |
75 scored questions across 12 realistic buyer and vendor journeys plus one-shot benchmarks, run on ChatGPT, Google AI Mode and Perplexity from a New Zealand context, three repetitions inside a 14-day window. Every conversation logged verbatim with citations; every factual claim about your firm verdicted against primary sources; scores computed under a published rubric. Sessions run in clean accounts with scripted phrasing — a deliberately repeatable lab condition — with a naturalistic-phrasing arm to test that rankings hold when queries get messy, the way real people type.
The full 75-question instrument, per-turn scoring rubric and every transcript are available in the evidence walkthrough.
A 30-minute walkthrough of your full evidence file: every conversation, every claim, every score — and exactly what we'd fix first. No deck, no obligation. Your scores move weekly whether you act or not; the walkthrough at least shows you what the AI is saying while you decide.
Get your findings — NZ$2,000 +GST