Customers no longer scroll through search results. They ask ChatGPT, Gemini and Perplexity for a recommendation, and they contact whoever the answer names. We fix the listings, pages and reviews those assistants read, then measure how often you get named.
Share of runs where the business is named · 20 prompts × 5 runs per engine
Illustrative example. Business names and figures are for demonstration only.
A results page showed ten businesses. An AI answer names three to five, and most people call one of them. If you're not in the answer, the customer never sees you.
Ten blue links, a map pack and ads. Ranking on page one was enough to get a call.
The assistant runs several searches, reads listings and reviews, and returns a shortlist. It favours businesses whose details it can confirm across several sources.
AI assistants don't recall local businesses from memory. They look them up live, the same way every time, and each stage has its own fix.
Each assistant pulls candidates from its own sources: map and places databases, review platforms and search indexes. If you're missing or inconsistent there, you aren't considered.
The assistant splits the question into several narrower searches. The pages that answer them, including local "best of" lists, directories and your own site, decide who gets pulled in.
From what it found, the model picks a shortlist. It weighs rating, volume, recency and what reviews actually say. A review that describes the specific job done is worth more than "great service!"
Almost none of the work happens inside ChatGPT. It's careful, unglamorous work on the places each assistant gets its information.
We claim and complete your profiles on Google, Bing Places, Apple Business Connect, Yelp and Foursquare, then fix stale citations elsewhere. Your name, address, hours and services end up identical in every source.
One page per service and area, each opening with a direct answer: response time, service radius, price ranges, licensing and after-hours policy. These are the specifics assistants quote. Schema markup is included.
We log every source an assistant cites for your target questions. Where a directory, local article or "best of" list recommends your competitors but not you, we work to get you on it.
Automatic review requests after every job, spread across the platforms each assistant reads. We ask customers to describe what was fixed and how fast. We never write reviews or offer incentives for them.
We make sure AI crawlers aren't blocked by your robots.txt or CDN (many sites block them by accident). We also set up Bing Webmaster Tools and IndexNow so changes are picked up within days.
Every month we re-run your full prompt set across every engine. You see your mention rate, your position and the sources behind each answer. That's how we tell which change caused which result.
AI answers change from one run to the next. Asking once and taking a screenshot is luck, not measurement. We run every target question several times on each engine and report what share of those runs name you.
| Cited source | Cited in | You |
|---|---|---|
| Review platform · category page | 46% | Listed |
| Industry directory | 31% | Outdated |
| Local news · "best of" feature | 18% | Missing |
| Places listing | 15% | Claimed |
| Third-party roundup article | 12% | Missing |
Illustrative example.
Most of the work happens up front. After that it's a monthly cycle: measure, find the gap, close it, measure again.
A baseline across all four engines, plus a full review of your listings and site.
Claim and correct listings, fix inconsistencies, unblock crawlers.
Build service pages and close the citation gaps the audit found.
Automatic requests after every job, and a reply to every review.
A report showing your mention rate, position and the sources that changed.
It pays off fastest when a single booked job is worth hundreds or thousands of dollars, and customers need someone now.
AI assistants change where they get information without warning. Anyone guaranteeing you "#1 in ChatGPT" is guessing.
Partly, and we'll say so. Much of the work is good local-SEO practice. The difference is where it's aimed. Each assistant reads a different mix of sources: ChatGPT leans on places data, Yelp and Bing; Gemini on Google; Siri on Apple Maps. So a business can rank well on Google and still be invisible in ChatGPT. We work on the sources each assistant actually cites, and we measure the results in the assistants themselves.
This is the most common case we see. Strong Google reviews don't carry over automatically. If your Yelp, Foursquare or Bing Places listing is unclaimed, thin or out of date, several assistants may barely know you exist. The audit shows exactly where the gap is.
Listing fixes typically show up within two to six weeks. Page and citation work takes one to two months. Reviews build up over several months and keep compounding. We re-measure at weeks 4, 8 and 12.
We run a set of real customer questions for your trade and area across ChatGPT, Gemini, Perplexity and Claude, several times each. You get your current mention rate, the competitors being recommended instead of you, and the specific sources behind those answers.
For the service pages and schema, yes, or we can work with your web developer. Listing and review work only needs access to the profiles themselves, and they stay in your name.
Book a call and we'll run your free AI visibility audit: your mention rate across four assistants, who's being recommended instead, and why.
Get my free audit