Local AI visibility audit: a checklist for one location
A practical checklist for auditing how AI search treats one business location: the prompts, the engines, the location setting, what answers cite, crawler access to the location page, AI-referred visits and the competitors named instead.
Multi-location reports are where local AI visibility gets confusing. Start with one location. If you can audit a single store or office properly, you can repeat it for the rest.
This is the order we work through, with the Promptwatch feature that does each step and the plan it needs. Plans matter here because several local features are gated.
1. Write down the location's facts
Before you open any tool, note:
- the exact address, and the name customers know the location by
- the category in plain words ("emergency dentist", not "oral healthcare provider")
- where customers come from: walk-ins from the street, short drives from nearby suburbs, or a service area you travel to
- the competitors you think you're up against
The last item is a guess. The audit will test it.
2. Build the prompt set
Aim for 15 to 30 prompts, which is the range Promptwatch itself suggests per monitor:
- discovery prompts without your name, such as "emergency dentist near me" and "dentist open Saturday near me"
- the same needs phrased as a longer question, without "near me"
- branded checks: "is [name] open on Saturdays", "does [name] take walk-ins"
- one or two comparisons with the competitors from step 1
Keep the city name out of the prompt text. The monitor sets the location. Label each prompt with Promptwatch's prompt types (organic, brand, competitor comparison) so you can filter later. If you're short of ideas, Prompt Explorer suggests prompts from Search Console, ranking keywords, People Also Ask or any page, including your location page.
3. Choose the engines and know how each gets a location
Pick the engines your customers plausibly use, then check how each one receives its location:
- ChatGPT, Google AI Overviews and Google AI Mode can answer from exact coordinates.
- Perplexity, Gemini, Copilot, Alexa and Meta AI use the monitor's country, state and city.
For one location that usually means two monitors: one on coordinates for the first group and one on the city for the second. Our guide to tracking Perplexity, Gemini and Copilot locally explains how to keep them comparable.
4. Set the location
- Coordinates: pin the monitor to the point that represents your customers and choose a radius between 20 m and 5 km. See how to choose a radius. The point can't be changed after the monitor is created.
- City: set the city that contains the pin. The state field is for US monitors only.
- Language: the language customers actually ask in.
Coordinates are on Business ($579/mo), Agency Scale ($799/mo) and Enterprise. City tracking is on Business and every agency plan from Kick-off ($199/mo).
5. Run a baseline
Two weeks of daily runs before any conclusions, and no edits to the monitor in that window. Changing models or prompts mid-baseline leaves you with a chart that mixes two setups.
6. Check what the answers say
Open the Responses tab and read a sample of answers for each engine. For each one, ask:
- Is the location named, and at what position?
- Is anything wrong: hours, address, a service you don't offer, a branch that closed?
- Did the engine actually answer for the area, or return a generic list?
Promptwatch scores visibility, Position in Answer and sentiment for you. Factual errors still need a human to read the answer.
7. Check what the answers cite
Open the Citations view for the monitor and note:
- whether your location page is cited, only your homepage, or nothing of yours
- which directories, review sites, local news sites and Reddit threads appear
- offsite mentions, meaning pages that name you without linking, which Promptwatch tracks separately
Citation Trends classifies each citation by content type and source type (directories are a source type) and shows when a page was first cited, its peak citation rate and how quickly that decays. A local answer built mostly on directories points at listings work. One built on a competitor's location page points at your own page.
8. Check crawler access to the location page
An engine that can't fetch the page has a hard time citing it. Promptwatch's Agent Analytics logs AI crawler visits, IP-verified so spoofed user agents are filtered out. For the location page, check:
- whether OAI-SearchBot, ChatGPT-User, PerplexityBot, Perplexity-User, Claude-SearchBot and Google-Agent have visited it
- any errors, with Promptwatch's explanation of each failure
- the crawl-to-citation path and the page's citation rate
Crawler logs start on Professional ($245/mo) and are on every agency plan. Essential has none. Logs come in through Cloudflare, AWS CloudFront, Fastly, Vercel, Netlify, Akamai, Google Cloud CDN, a custom HTTP endpoint or the WordPress plugin. The weekly site audit adds crawlability checks on top.
9. Check AI-referred visits
Install the visitor analytics script, the Google Tag Manager template or the WordPress plugin. Then look at:
- AI-referred visits to the location page, split by platform
- Top Locations, to see whether AI-referred visitors are actually near the location
- conversions from those visits
A location can be named often and send almost nobody, or be named rarely and convert well. Those two cases call for different next steps, which is why this check sits next to the visibility numbers.
10. Check who is named instead
Compare the businesses AI names with your list from step 1. Promptwatch auto-discovers competitors from the answers and shows their visibility, share of voice and position, and competitor heatmaps put them against topics or models. Brand merging combines duplicate entries for the same business so its counts don't split across spellings.
If the names differ from your list, you've found the businesses whose cited sources to study first.
11. Turn findings into actions
Write each finding as a single action with an owner. Promptwatch's Action Board holds custom actions with assignees and comments, and its Discovery agent reviews the data weekly and suggests up to five evidence-backed actions per run. After a month, re-run this checklist on the same monitors and compare.
Which plan covers the whole checklist
Everything above runs on one Promptwatch Business plan: coordinates, city tracking, citations, crawler logs (100M on Business) and visitor analytics. Agencies need Agency Scale for coordinates. Kick-off and Growth cover the rest of the list at city level. Before scaling the audit to many locations, see what city-level AI tracking costs per city, and for the full set-up of a pinned monitor, our coordinate tracking guide.