Track LLM Visibility

More and more customers ask an AI assistant for a local recommendation before they ever open Google. LLM Visibility shows you whether ChatGPT and Gemini name and cite your business in the responses they give to the prompts your customers actually use.

What it tracks and why it matters

LLM Visibility watches the two assistants your customers increasingly use to find local providers: ChatGPT and Gemini. When someone asks "who's the best plumber in Manchester?", each returns a written response rather than a list of links. This feature tells you whether your business gets named, recommended or cited in that response, and how often you show up against everyone else.

Each engine has its own results, because they do not answer the same way. ChatGPT often names a shortlist of local businesses in its answer, which gives the fullest competitor picture. Gemini grounds its answers in Google's own index, which makes it the closest read on how the work you already do on your Google Business Profile, citations and content shows up in an AI answer. ChatGPT does not read your Google Business Profile directly, so being strong on Google does not guarantee you are named there. Both engines are asked the same prompts for the location you are measuring, so the two can be compared like for like.

This matters because a recommendation inside an assistant often replaces the click that used to land on a search results page. If you are invisible there, you lose the customer before they reach your website or your Google listing.

How your keywords become prompts

You do not have to guess what people type into an assistant. SearchOps groups the keywords you track at a location into the 3 to 15 services you actually sell, then asks prompts for each service the way real customers ask: a plain ask, plus one for each customer you measure, such as someone watching the cost or someone who needs help tonight. Twenty ways of searching for "boiler repair" become one service, asked about once, so the cost depends on how many services and customers you measure, not on how many keywords you keep.

You can see and change all of it in the location's Keywords page (Track > Keywords), service by service: the services, their prompts, your customers and how many prompts each service gets. See Set up your services, customers and prompts.

What the results show

Because these assistants answer in prose rather than an ordered list, there is no position to report, and we never make one up. Instead you get measures built for how responses actually work:

• Visibility rate: how often you are named or cited across the prompts put to each engine, reported per engine so you can see where you are strong and where you are missing.
• Share of voice: how often you show up compared with the other providers named for the same prompts, alongside the competitors each engine names.

Each response is also sorted by how you featured: Shortlisted (named in the engine's list of options), Mentioned, Cited only (your site was used as a source but you were not named), or Absent. Where the engine writes a line or two about each business it lists, a response that wrote one about you is marked Named with reasons, because being listed with a reason to choose you is a stronger result. If you have several locations, a response that names a different one of them is marked Another of your locations rather than counted as a win.

You get a trend over time rather than a single snapshot, the full response saved word for word so you can read exactly what was said and who was named, and, where the engine cites its sources, the pages it trusted so you know what to strengthen.

The two engines will not always agree, and that is useful. Being named by one and missed by the other tells you the gap is with that engine's sources rather than with your business.

Where to find everything

Open Track > LLM Visibility in the sidebar. Across the top of the section are its own tabs: Overview (your locations), Insights, AI Citations, Keywords and Schedules.

Open a location to read its latest check. Its tabs are:

• Results: how often ChatGPT and Gemini named you in the latest check, side by side, then a grid of your services against your kinds of customer showing where you were named (click a square to read those responses), and the businesses named instead of you.
• Responses: every response saved word for word, filtered by outcome, prompt group, stage and who was asking, with an option to compare against the previous check.
• Brand: what the engines say about you once someone asks about you by name, and what they have to work with about your business. The prompts themselves are edited under Your name on the location's Keywords page, where you can also add up to 5 of your own.
• Prompts (on the right) takes you to the location's Keywords page, where its prompts, services and customers are edited.

Insights puts the two engines side by side over 7, 30 or 90 days or all time, with tabs for Competitors, Keywords, Sources, Who's asking (each customer compared with the plain ask), Over time (each engine check by check, and visits AI answers sent to your website) and ChatGPT's own searches (the web searches ChatGPT made before answering; Gemini does not report its searches). See Read your LLM Visibility Insights.

AI Citations shows which websites the answers relied on. See See which sites AI answers rely on.

Run a check

Most people set a schedule once and come back to read the results. To run a check yourself, press Run a check at the top of a location, on its row in the Overview, or in its Keywords page. The dialog shows what the check will cost, service by service, and how many checks you have left before anything runs. Both engines are ticked; checking one engine halves the cost, but that check can't be compared side by side in Insights. Your current results stay on screen until the new ones arrive.

Beside Run a check is the location's schedule, which reads Not scheduled until you set one. Open it to run checks weekly, every 2 weeks or monthly, on the day and time you pick, with roughly how many checks a month that uses. Nothing is scheduled until you press Save, and SearchOps never sets one up for you.

How it differs from geo-grid scans

LLM Visibility and geo-grid scans do different jobs, and you want both.

• Geo-grid scans cover Google's own surfaces from a grid of points around your business: your Google Maps position, whether you are in the local pack's top 3 and cited in AI Overviews, and whether Google's AI Mode names you. See Choose what a scan measures.
• LLM Visibility covers the assistants themselves, ChatGPT and Gemini, for your town or city, and reports a visibility rate and share of voice, not a rank. A Gemini-powered AI Overview sits on the Google results page, while the Gemini assistant is somewhere a customer goes directly.

Availability and cost

LLM Visibility is included on every paid plan, in the sidebar under Track. Each plan comes with a monthly allowance of AI visibility checks: 600 on Starter, 3,000 on Pro and 6,000 on Agency, with Enterprise set by contract. One check is one prompt put to one engine once, so a prompt asked of both engines is two checks. Checks you run yourself and scheduled checks both ask each prompt once. Schedules set up before 17 September 2026 ask each prompt three times, which the schedule's page shows as "× 3 samples". Your allowance resets with your plan each month. See How credits and quotas work.

Next steps

• Shape what is asked: Set up your services, customers and prompts.
• Spot what keeps happening across checks: Read your LLM Visibility Insights.
• Add the customers you actually get: Ask as the customers you actually get.
• Find prompts your own data points to: Research your prompts.