Methodology

How we test what AI says about you

Same questions, same tools, logged the same way every time, so the numbers mean something and you can compare month to month.

1. We write the questions homeowners actually ask

We build a fixed set of questions for your city and services, phrased the way homeowners type or speak them. For example:

• "Who's the best kitchen remodeler in [your city]?"
• "Bathroom remodeling companies near [your neighborhood] with good reviews"
• "How much does a kitchen remodel cost in [your city], and who should I call?"

We include broad questions, service-specific questions and price-driven questions, because each one pulls a different set of names.

2. We run them across four tools

Each question goes to ChatGPT, Gemini, Perplexity and Google (including its AI answers where they appear). We use fresh sessions so personal history doesn't skew the result.

3. We log and count every answer

Every company named gets recorded. Your result is a simple count, such as "named in 2 of 52 answers", next to the counts for the competitors who showed up most. Where a tool cites sources, we record those too.

4. We trace the sources

AI tools don't invent recommendations from nothing. They lean on Google Business Profiles, review platforms, directories, "best of" articles and company websites. We check which of those sources mention your competitors and not you.

5. We check your foundations

Google Business Profile categories and completeness, review count and recency, name/address/phone consistency, duplicate listings, service pages and structured data on your site.

6. We rank the fixes

Every finding is rated High, Medium or Low, with what we found, why it matters, how to fix it and roughly how long it takes.

What we don't do

• We don't promise rankings or placements. AI tools change, and nobody controls them.
• We don't post fake reviews or buy links.
• We don't cherry-pick. If a tool named you once and ignored you nine times, the report says 1 of 10.

AI answers vary from one run to the next. That's why we use a fixed question set and repeat it, rather than relying on one screenshot.