A national AI tracker can't see your client's business.
When someone asks ChatGPT for a plumber, a dentist or an injury lawyer, the answer depends on where they are standing. Cite AI asks from inside each market your clients serve — so you can show a roofer in one metro exactly who the AI recommends instead of them, and what to do about it.
10 client projects · 500 tracked prompts · about $30 per client
Panels that run in your client’s market
Tag a prompt with the city or metro a client serves and we ask each engine as a searcher standing there. The businesses that come back are the ones competing for that client’s calls — not the national chains a location-less tool returns. For multi-location clients, switch the dashboard between their markets one at a time.
A map of the metro, not one number for it
Switch grid tracking on and a prompt runs at nine or twenty-five points across the client’s market instead of once. You get the neighbourhoods where the AI names them and the ones where it names a competitor — the difference a single city-wide reading hides completely. a 3 × 3 counts as four tracked prompts and a 5 × 5 as twelve, against the same pool.
Receipts for the monthly report
Every citation comes with the exact sentence the AI wrote and the page behind it. Every miss shows whether the client’s page was read and skipped or never surfaced at all. That is a slide with evidence on it, not a number you are asking the client to trust.
It says which lever to pull
Written answers move with content and structure. The map card inside the AI app is Google’s listing data passed through, so that one is a Google Business Profile job. We measure it at Google’s own local pack — your client’s rank at each point of the grid — because that is the number that moves when the profile improves.
One project per client
Each client gets its own prompt panel, its own markets, its own competitors, reports and audit history. Month-end reporting is per-client by default and nothing crosses between accounts.
One reading per metro hides the answer
Ask for a plumber from uptown and from the suburbs and you get different businesses — the answer moves with where the searcher is standing. Grid tracking measures a prompt at every point of a grid across your market, so you see the neighbourhoods where your client is named and the ones where they are not, rather than one number for the whole city.
Real places, not coordinates
You set the spread — a 3-mile radius for a dense city core, up to 25 for a client covering a whole metro. Every point resolves to a named neighbourhood or town before it runs, so a result reads “eighth in Pineville”, not “eighth at 35.086, −80.892”. Points landing on water or open farmland are dropped and the map says how many; if too few resolve, we tell you on save and the prompt keeps running as a single city-level reading.
Named and ranked are different facts
The map separates the points where an AI actually named your client from the points where Google ranks them but the AI passed over them. It also separates “ranked below 20” from “not in Google’s list here” — the first is a profile you improve, the second is usually a service area or category that does not cover that point.
Priced below what it runs
A 3 × 3 runs at 9 points and costs 4 prompts from your pool; a 5 × 5 runs at 25 and costs 12. No separate grid allowance to buy — it comes out of the same pool.
Grid readings come from the two engines that take a set of coordinates — ChatGPT and Gemini — plus Google's own map results at each point. We measured that it matters: twenty-one different businesses came back across a nine-point grid and eight of them appeared at only one point. The table below is the full picture, including the engines a local project deliberately does not run.
Two surfaces, two levers — and one of them isn't yours.
A local AI answer is really two things stitched together: a written recommendation citing websites, and a map card drawn from Google's business listings. You can move the first with the content work you already sell. The second moves with the client's Google Business Profile. Any tool that blurs the two is setting you up to promise a result you can't deliver.
Google rank gets your client into the running
Across 5,130 ranked businesses, a business in Google’s top three was named by ChatGPT 48% of the time and by Gemini 76%. Ranked 51st or lower, both named it 1% of the time. Local rank is not a separate game from AI visibility — it is the gate in front of it. If your client is not ranking, that is the first job, and it is the one most agencies already know how to do.
Ranking is not enough, and the gap is what we measure
Even in Google’s top three, ChatGPT skips the business more than half the time. That gap is the whole product: we show you which points, which engines and which competitors got named instead — so the conversation moves from “we are ranking” to “we are ranking and still not being recommended, here is where”.
The two engines do not behave the same way
When the map fires, Gemini answers purely from Google’s listings and cites no web pages at all. ChatGPT runs both at once — it reads the map and the web in the same answer, and puts about ten businesses on its map card while writing only two to four into the prose a customer actually reads. Tracking one engine tells you about that engine.
| Engine | Local precision | What we can tell you locally | Lever |
|---|---|---|---|
| ChatGPT | Point-levelMeasured at every grid point | Which businesses ChatGPT names, and which sites it cites, for a searcher standing at a specific set of coordinates. Run at every point of a grid, nine or twenty-five readings across one metro rather than one. | Website content and structure for the citations; the Google Business Profile for the businesses it names off the map. |
| Google AI Overviews | City-level | The AI answer at the top of Google, captured once from within the market you name — the same answer a searcher standing there gets. (Google’s map pack is read separately at every grid point; that is where your ranked position comes from, and it counts under the map surface rather than under this engine.) | Website content and structure; the local pack alongside it is driven by the Google Business Profile. |
| Google AI Mode | City-level | Google’s conversational search, run against a named location rather than a generic national query. | Website content and structure. |
| Google Gemini | Point-levelMeasured at every grid point | Two surfaces, measured differently. The businesses Gemini surfaces from Google’s listings are read at every point of the grid — its Maps grounding takes coordinates, and in our nine-point test across one metro it named a substantially different set at each. The websites it cites are read once for the market: we have not measured whether that layer varies inside a metro, so we do not grid it. | Google Business Profile, and reviews specifically. Holding Google rank fixed at 1-10, Gemini named 29% of businesses under 50 reviews and 75% of those past 200, and the pattern held in crowded and thin markets alike. It is the one lever we have measured moving an engine. |
| Claude | Not run locally | Nothing local. Claude’s API returns cited links and no place data — the map results you see in the Claude app come from Google Places, which the API does not expose. We drop it from local projects rather than bill you for a reading that cannot see a place. It runs in full on national projects. | Website content and structure, at national scope. |
| Perplexity | Not run locally | Nothing local. Perplexity has no map surface, and its location signal did not separate one metro from another in our testing — so we drop it from local projects rather than present national results as local ones. It runs in full on national projects. | Website content and structure, at national scope. |
A local project measures four of the six engines. Claude's API returns no place data, and in our own testing Perplexity's answers didn't differ between one metro and the next — so neither runs on local projects. Both run in full on national ones. Where an answer draws on Google's business listings rather than the open web, the lever is your client's Google Business Profile, not their website.
The monthly client workflow
- 1
Add the client as a project with the city or metro they serve. Cite AI drafts a prompt panel the way their customers actually ask — “near me”, “open now”, “best X in [city]” — or you write your own.
- 2
Attach the client’s Google Business Profile. AI answers name local businesses without linking to them, so we match your client by their Google listing rather than by their website. Search the name, pick the profile — about a minute per client, and it is what the map matches on.
- 3
Turn grid tracking on for the project: pick 3×3 or 5×5 and a radius from 3 to 25 miles. We place a grid of real places across that area — named neighbourhoods and towns, never bare coordinates. Then flag the individual prompts you want measured across it; nothing fans out until you do.
- 4
Every week we ask each engine those questions from each of those locations and record which businesses got cited, which got named, where they ranked on the map, and how the client was described.
- 5
Run the site audit. The fix briefs point at specific pages and specific gaps, so the work is scoped before you quote it — and results that trace to the map surface get flagged as GBP work instead.
- 6
Report the movement with quoted snippets as evidence: who was cited last month, who is cited now, and what changed after you shipped.
Agency plan
- 500 tracked prompts
- 10 active client projects
- Per-market prompt targeting
- 6 engines nationally · 4 on local projects, 2 at every grid point
- Per-client reporting and history
- Site audits with advanced page limits
- Content fix briefs
- Citation-level receipts
Sizing the pool: each prompt targets one market, so a question tracked across three metros is three prompts. Around 50 gives a single-market client a stable baseline; multi-market accounts need a multiple of that, which is why the 500 is a shared pool rather than a per-client cap.
Agency questions
How does per-client pricing work?
One Agency subscription ($299/month) covers 10 active projects and a shared pool of 500 tracked prompts. You allocate the pool across clients — evenly at 50 each, or weighted toward the accounts that pay for it.
I only run local clients. Am I paying for two engines I don’t get?
No — same price, same prompt count. What changes is where the spend goes. On a local project the two engines that cannot see a place are replaced by depth: ChatGPT and Gemini are each measured at every point of the grid rather than once, and Google’s map pack is read at every point too. One prompt at a 5×5 grid is fifty AI readings — the two gridded engines at twenty-five points — plus a map-pack read at each of those points, where a national project gets six readings in total.
Does one client need more than one project?
Usually no. One project per client, set to local, covers the map, the grid and the citation tracking. You would only add a second if you want a national baseline alongside the local panel, because market reach and business identity are set per project. Ten projects covers ten clients — talk to us if you need more.
Does a grid eat my prompt pool?
It costs more than an ordinary prompt, because it really is more measurements: a 3 × 3 counts as four tracked prompts and a 5 × 5 as twelve — far below the nine or twenty-five readings it actually runs, but not free. There is no separate grid allowance to buy: it comes out of the same pool. Grid tracking is on Starter, Growth, Pro and Agency.
How many points should I use?
Three by three for a single-location client — nine readings, centre plus the ring around it, enough to see whether the AI’s answer holds outside their own postcode. Five by five for a client whose service area is the whole metro, or when a 3×3 comes back uniform and you want to know where the edge is. Keep the points at least five miles apart: below that, neighbouring points tend to return the same answer and you are paying for repetition — widen the radius or drop to 3×3. You get one map per gridded prompt, so grid the two or three prompts that drive calls rather than the whole panel.
What if a client serves several cities?
Each prompt targets one market, so a question you want tracked in three metros is three prompts from the pool. That is the honest arithmetic: a roofer covering a metro and two suburbs needs a bigger share of the pool than a single-location dentist. Around 50 prompts gives a single-market client a stable baseline, and a multi-market client needs a multiple of that — which is why the pool is shared rather than split evenly.
Does this replace the Google Business Profile work I already do?
No, and we will not pretend otherwise. The map results inside ChatGPT and Gemini come from Google’s business listings, so GBP remains the lever there — we measure that surface rather than claim to move it. What that measurement is worth to you is the map view: it shows the client’s pack rank at every point of their market, so a GBP job stops being “improve the profile” and becomes “we are eighth in the three suburbs where you are not getting calls”. The written answer is the other half, where the AI cites websites — that one responds to content and structure, which is billable work you can now show a before and after for.
Can I show the results to clients?
Yes. Each project has its own reports, citation history with quoted snippets, share-of-voice trends and audit findings, built to be screenshotted or exported into your own deliverables.
Do you offer white-label reporting?
Not yet — reporting is Cite AI branded, and you export or screenshot it into your own client deliverables. If white-labeling is what stands between you and reselling this, tell us at info@usecite.ai; that feedback shapes what we build next.
What happens when a client wants a bigger panel?
Reallocate prompts from the shared pool, or contact us about a custom plan (1,000+ prompts, custom project counts, daily update schedules).
Run it on one client this week.
Pick the client who asks about ChatGPT most. Add their city, let it draft the panel, and see who the AI recommends in their market instead of them.
14-day free trial · No credit card required · 20 prompts & 1 project limit