45%
of consumers now ask AI for local recommendations
up from 6% a year earlier
Method
Being named is buildable. Most of it is not on your website.
Six phases, one score. Published in full because the difficulty is in doing it consistently, not in knowing what it is.
Why one score is not enough
Four engines, four different levers
| Engine | Reads |
|---|---|
| ChatGPT | Bing · licensed publishers |
| Gemini / AI Overviews | Google index · entity graph |
| Perplexity | Own index · Reddit |
| Copilot | Bing · editorial |
An agency working only on Google is working on one surface out of four. We report each engine separately because the fix for one is rarely the fix for another.
6.5×
more likely to be cited via third-party sources than your own domain
the reason most of this work is off-site
30+
reviews at 4.3 stars — the observed entry point
100+ in competitive metros
1–5
businesses named in a typical AI answer
there is no page two
Sequenced by score impact
The six phases
-
Baseline
Fifty prompts your customers actually use, run across all four engines, three times each. For every prompt we record whether you were named, who was named instead, in what position, and which source the engine cited. That last column is the diagnosis.
-
Entity foundation
Make the business machine-readable. Structured data for the organisation, practitioners and services. Name, address and phone consistency across Business Profile, Bing Places, Apple Business Connect, Foursquare, Yelp and the directories specific to your category. Confirm no AI crawler is blocked in robots.txt — a surprising number of agency-built sites block them by accident.
-
Review velocity
The least glamorous phase and the one carrying the most weight. Observed practice puts the entry point near a credible volume of independent coverage, and above a hundred in competitive metros. Recency counts as heavily as volume — two hundred reviews with nothing in six months loses to sixty updated weekly.
-
Third-party presence
The phase most agencies skip entirely, and where the six-times multiplier lives. Local editorial coverage, genuine participation in the communities your customers actually read, practitioner video, and the vertical authority profiles for your category. Never astroturfing — it is detectable and it backfires.
-
Extractable content
Now the website, built for extraction rather than for browsing. Named practitioners with credentials, direct answers under headings phrased the way people ask, real specifics, published pricing, visible update dates. Research on generative engines consistently finds citations and statistics lift visibility, and keyword stuffing actively reduces it.
-
Convert
Visibility that does not book is a vanity metric. Analytics segmented by AI referrer, booking paths on every service page, call tracking with source attribution, and speed-to-lead on inbound enquiries.
Reported monthly
What we report
- Named Score — share of tracked prompts where you are named
- Per-engine breakdown — because the four move independently
- Share of AI voice — against three named local competitors
- Citation source map — which source produced every mention
- Review velocity — against target, per location
- Booked appointments — attributable to organic and AI referral
What we cannot tell you
No AI platform publishes how it selects sources. Everything above is measured output plus correlations documented across the industry — not confirmed ranking factors.
The signals also move. Community content's share of citations shifted by double digits in a single quarter, and the most-cited social platform changed inside a year. We re-baseline the source map quarterly, which is the honest reason this is a retainer and not a one-off project.
Anyone guaranteeing you a placement in an AI answer is either misinformed or selling you something. We report the number, honestly, whether it moved or not.