PlumberFindable field guide
Do Reviews and Ratings Shape the Answers AI Gives About Your Plumbing Company?
Six review signals AI systems actually use
1. Review velocity and recency
A steady flow of recent reviews signals an active, in-demand business. A profile that went quiet two years ago reads as stale, no matter how good the old ratings were.
2. Rating distribution, not just the average
A 4.8 across 300 reviews beats a perfect 5.0 across 12. AI systems treat thin perfect scores as low-confidence evidence and favor profiles with a believable spread.
3. Owner responses with specifics
Responding to negative reviews with specifics (the date, the job type, what was done to fix it) is quotable evidence of accountability. Silence is quotable too, and not in a good way.
4. Review text specificity
Reviews that name the actual service, like water heater replacement or sewer camera inspection, help an AI assistant match you to the exact question a customer asked. Generic praise does not.
5. Source diversity
Google, Yelp, Angi and BBB mentions across the web build entity confidence. A business that exists on only one platform is thinner evidence, and thinner evidence gets cited less.
6. Negative review patterns
Repeated complaints about the same thing, no-shows, pricing surprises, messy job sites, become the cons section of an AI-generated comparison. One bad review is noise; a pattern is a verdict.
Review profiles that get recommended vs ones that get skipped
| Gets recommended | Gets skipped |
|---|---|
| 200+ reviews across three or more platforms, with fresh ones every month | Nine reviews, all from two years ago, all five stars, all suspiciously similar |
| Owner responds to every review from the last year, including the bad ones | No responses at all, or a copy-pasted thanks on every positive review |
| Reviews name specific jobs: repipes, sump pumps, gas line work | Every review says great service with no detail about what was actually done |
| Photos in reviews showing trucks, uniforms, finished work | Text only, no photos, no location tags, no evidence |
| A believable 4.6 to 4.9 with a few critical reviews that were answered well | A flat 5.0 that looks filtered, which reads as less trustworthy, not more |
Frequently asked questions
Do AI assistants actually read Google reviews?
They read the data around them: review counts, ratings, recency and response patterns from business profiles, plus review text quoted on sites across the web. A review strategy is now an AI visibility strategy.
How many reviews are enough?
There is no magic number. What matters is the competitive set in your market. In most metro areas, 100 or more reviews across platforms puts you in the game; fewer than 30 usually leaves you out of AI-generated shortlists.
Do responses to bad reviews really matter for AI?
Yes. A thoughtful response to a one-star review is one of the few places a business gets to put its side of the story in quotable text. AI systems can and do cite it.
Should you ask every customer for a review?
Ask everyone equally and make it easy with a direct link on the invoice. Never gate requests by asking only happy customers, and never offer incentives. Both distort the profile and violate platform rules.
Do photos in reviews help?
They do. Customer photos of your trucks, your techs and finished work are hard-to-fake evidence. They also make your profile more likely to be the one an assistant describes.
What is the fastest review fix a plumber can make this week?
Claim and complete every profile, respond to every review from the last twelve months starting with the negative ones, and add a review link to every invoice and follow-up text.
Get the free audit
Receive a prioritized review of your discovery, answer, trust and action gaps.
Request the free audit