How to monitor local competitors in AI answers
Your competitor in one city is often not your competitor in the next. How to track local rivals in ChatGPT, Gemini and Google AI answers with per-city and coordinate monitors, auto-discovered competitors, brand merging and agent-assisted gap analysis.
Ask a national competitor list who you lose to in AI answers and you'll get the same five names everywhere. Ask city by city and the list changes. In one city it's a national chain. In the next it's an independent with one location and a well-cited website. In a third it's a directory page that ranks providers and names you nowhere.
So local competitor monitoring has to start from the city answer, not from the list in your head. This guide shows how to set that up in Promptwatch and how to use an agent to turn what you find into a gap list.
Track each place separately
Competitor numbers are only as local as the monitor behind them. In Promptwatch a monitor carries a location, a language and the models to run, and its prompts run daily per model. For local work, create one monitor per location with the same prompts in each, so a competitor's share in one city compares fairly with its share in another.
City and state tracking start on Business ($579/mo) and are on every agency plan, from Kick-off ($199/mo). Essential ($95/mo) and Professional ($245/mo) stop at country level.
For a business that competes block by block, such as a dense city with several branches, Business, Agency Scale ($799/mo) and Enterprise add local GEO/SEO coordinate targeting. That pins a monitor to exact coordinates with a radius from 20 m to 5 km. ChatGPT, Google AI Overviews and Google AI Mode answer as if asked from that spot, while other models fall back to country, state or city. Two coordinate monitors a few kilometres apart can show different rivals in those three engines. Whether they do in your category is worth checking before you build a whole grid.
Let the tool find the rivals you didn't list
The first useful thing a local monitor tells you is who you're up against. Promptwatch auto-discovers competitors, so the independent in one suburb can show up even if nobody on your team has heard of it.
Then clean the list. Local brands often appear under several names: the trading name, the legal name, and a "Brand Leeds" form for each branch. Brand merging combines those into one competitor so its share isn't split across three rows. Do the same for yourself with brand aliases, and use the Brand Book's analysis instructions to set what counts as a mention of you. If one branch trades under a slightly different name, this is where you tell the platform it's still you.
Skip this step and your share of voice per city will be wrong in ways that are hard to spot later.
Read the competitor numbers per city
For each city monitor, Promptwatch reports competitor visibility, Share of Voice and Position in Answer. Read them together:
- Visibility tells you how often a rival appears in that city's answers.
- Share of Voice tells you how much of the answer it takes compared with everyone else.
- Position tells you whether it's named first or mentioned at the end.
A rival that appears often but sits low in the answer is a different threat from one that appears less often but is always named first. The A/B period comparison shows whether either moved this month.
Competitor heatmaps help when you have many rivals and many models. A heatmap compares any two dimensions, such as brands against models or brands against topics, so you can see that a rival is strong on ChatGPT but rarely named by Perplexity. For a brand in several countries, brands against countries works too. A custom dashboard built from the competitor comparison template keeps the city view in one place for whoever reviews it each week.
Find out why a rival wins
Share numbers tell you who wins. Citations tell you why. For each city where a rival leads, open the citation analytics for that monitor and look at which pages and domains the answers drew on. The sources tend to fall into a few groups: the rival's own city page, a local guide or directory, and community content such as a Reddit thread or a YouTube video.
Each points to a different response. If the rival's city page is cited, add that URL to Page Tracker, which follows specific URLs, including competitors', and shows when they're cited. If a directory is cited, check whether you're on it and whether your details there are right. The offsite mentions report covers third-party pages that name a brand without linking to it, which is where unlinked local press coverage turns up.
Don't try to reverse-engineer a ranking formula from this. Treat it as a list of sources the engines used in that city, and ask whether you belong on them.
Hand the gap analysis to an agent
Doing the above for 15 cities by hand is a week's work. Agent Chat does it in a conversation. It runs multi-step workflows over live data, can split work across read-only subagents, and draws charts. One of its Agent Skills is a competitor gap analysis.
A useful ask, for a hypothetical project:
"For each city monitor, name the competitor with the highest Share of Voice over the last 30 days. For each one, list the three domains most cited in answers where they appear and we don't. Flag any city where the top cited page is the competitor's own location page. Put the results in one table and a bar chart of our share against theirs per city."
Then take the gaps into content gap analysis, which scores how well your pages cover the prompts a city's answers respond to. If you want the comparison to arrive weekly, save the ask as a Scheduled Chat; our guide to automating local AI visibility reporting covers the setup.
What to do with the findings
Most local competitor findings end in one of a few places: a page that needs real local detail, a listing that's missing or wrong, or a source you could reasonably earn a mention on. Put each on the Action Board with an owner and the citation evidence attached. Our guide to the Action Board for local fixes walks through that.
Plenty of trackers show a competitor list. What makes local competitor monitoring useful is going from "this independent beats us in one suburb" to "because the answer cites its service page and a local guide we're missing" without leaving the tool. That chain of per-city monitors, merged competitors, page-level citations and an agent that writes the gap list is why we'd run it on Promptwatch, starting on Business for city tracking or any agency plan for client work.