Home > Methodology
How we measure what AI recommends
Customers increasingly ask AI assistants, not search engines, who to hire. AI+ Local makes those answers visible: who gets named, who doesn't, how much the assistants agree, and how it all changes over time. We don't sell placement and we don't editorialize the results. We measure what the assistants say, and we publish it.
The Process
- We ask the questions customers ask. Every week, each area and category gets the same consumer-style questions (“Who's the best plumber in South Tampa?” and variations by price and speed), put to ChatGPT, Gemini, Claude, Perplexity and Grok. Identical wording, every assistant, every week, so the answers compare cleanly across time.
- Always web-connected. Each assistant answers in its live, search-connected mode, the same one a customer gets, so the results reflect today's web and we can see which sources it leaned on.
- Names become canonical businesses. “Hiller,” “Happy Hiller” and “Hiller Plumbing LLC” are one company. Every name an assistant says is matched to a single real-world business, so counts and trends never double-count variants.
- Sources are first-class. Every page an assistant cites is recorded and classified: directories, review sites, local news, community threads, the business's own site. The citations are what explain a recommendation.
- Snapshots over time, never overwritten. Each run is a new entry in a time series; nothing is edited after the fact. The rankings reflect the full history: how often a business is named, how many assistants agree, how settled the answer is, and which way it's moving.
Reading the Numbers
The same terms appear across every area, category, and report
| AI rank | A business's place in the combined answer across all five assistants. #1 is the top recommendation in an area. |
| Named by | How many of the five assistants named the business, like 4 of 5. The closer to five, the more the assistants agree. |
| Change | Movement since the last run: 2 climbed two spots, 2 slipped two; “new” just entered, “out” left the rankings. |
| Share of voice | The share of all AI mentions in an area that one business holds, against everyone else named. |
| Rankings by intent | The same area asked four ways (best, most affordable, cheapest, fastest), sorted into the answer (#1), podium (#2–3), named (#4+), or not named. |
| Race | How settled the answer is week to week: stable consensus, competitive, or open race. |
| Sources | Every page the assistants cite, ranked by share of citations and how many assistants used it. |
The Pro Report
The public pages refresh weekly. A Pro Report tracks a single business and refreshes daily, sometimes more often, so an owner sees a change the day it happens instead of a week later.
It pulls the whole picture into one report: every area the business is tracked in, with each assistant's exact position and full rank history; how it ranks when buyers ask by value, price, and speed; its share of the AI conversation against the rivals named alongside it; the sources the assistants cite and which ones don't yet mention it; and a short, ordered list of where to focus next.
What We Don't Do
- We don't sell placement. Rankings come from the assistants' outputs alone.
- We don't cherry-pick runs. Every snapshot enters the record, including the ones where the assistants disagree or change their minds.
- We don't hide thin data. Areas where the assistants haven't named enough distinct businesses are marked as early rather than padded.
Questions about the data, or want your business tracked across every area it serves? Get the Pro Report.