AI Named SERVPRO 142 Times, 13 Locals 56

This is the fifth study in the series and the fourth in the same thirteen cities. In October we tested thirteen HVAC companies, thirteen plumbers and thirteen electricians that hold Google's first page in their metro. For restoration we kept the cities and the method: thirteen water damage restoration companies, each the first restoration company's own site on Google's first page for "water damage restoration [city]", in Boston, Atlanta, Seattle, Minneapolis, Columbus, Charlotte, Tampa, Dallas, Phoenix, Chicago, Nashville, Denver and Philadelphia.

The restoration companies were named in 56 of 260 answers, 22 percent, between the plumbers and the electricians. The franchise pattern from the plumbing and electrical studies did not just repeat; it took over. SERVPRO was named in 142 of the 260 answers, in all thirteen metros, two and a half times as often as the thirteen first-page local companies combined. Four more national names, Paul Davis, PuroClean, ServiceMaster and 911 Restoration, were each named more often than any local company except one.

13

restoration companies, one per metro, each on Google's first page for its city

56 of 260

answers named the tested restoration company, 22 percent

142

answers named SERVPRO, in all 13 metros

213

citations to servpro.com, the most-cited source, ahead of every directory and review site

3 of 13

companies were never named in any of their 20 answers

As before, this is not a judgement on anyone's drying, demolition or rebuild work. It is a measurement of what four assistants do with a local water damage question, and of what they build the answer from.

How we picked the 13 restoration companies and what we asked

We pulled Google's organic results for "water damage restoration [city]" from Ahrefs SERP data on November 10, 2026, in the same thirteen metros. In each we took the first restoration company's own site. We skipped directories and review sites, Reddit, Facebook, national and franchise brands, multi-metro chains, and any company already in an earlier study.

Restoration is the most franchised trade we have tested, and it showed in the selection. In Boston no local company's own site appeared in the first eleven organic results for either keyword we tried: the page is franchise pages, directories and marketplaces, so we took the local company Google's own AI Overview listed for the same query. In Nashville the first qualifying result turned out to be a branch of a company based in Longview, Texas, and the testing tool wrote all five questions about Longview, so it was excluded and the first local-pack company whose site the tool could read took its place. The first Atlanta result was a franchise location under a local-looking name and was excluded for the same reason. Tampa needed a keyword variant because the tool could not read the first qualifying site. The selection file in the data pack records every skip and every swap.

The five buyer questions per company follow the work each company advertises: emergency water extraction in Boston, flood cleanup in Nashville, commercial restoration in Columbus, mold and fire alongside water almost everywhere. Across the 65 prompts the shapes were:

1. 01Who are the top water damage restoration companies in [metro] in 2026? The plain list question, 40 of the 65 prompts in one form or another.
2. 02Which restoration company in [metro] do you recommend for [a loss]? A burst pipe, a flooded basement, storm damage, a sewage backup.
3. 03Where can I find an affordable water damage restoration company in [metro]? or Which offers the best value?
4. 04Which restoration company in [metro] has the best reviews? or is the most recommended?
5. 05Who does the best [job] in [metro]? Emergency water extraction, mold remediation after a leak, fire and smoke cleanup.

Two companies' prompts came back with no metro at all. For the Phoenix company the tool asked about "the United States"; for the Denver company it asked about no place whatsoever. That is a finding, not a flaw: the tool writes its questions from what the website says, and neither homepage names the market it works. Both companies were named in none of their twenty answers. They are kept in the 260 and flagged in the data pack.

Each prompt went to ChatGPT, Gemini, Claude and Perplexity through the same visibility-testing connector as the earlier studies: fresh sessions, no account history, no location set, the smaller tier of each assistant. Twenty answers per company, 260 in total, collected on November 10, 2026.

Two numbers per company, as before. Linked, of 5: in how many of the five prompts did at least one assistant link the company's own website. Named, of 20: in how many of the twenty answers did the company's name appear at all. And, as in the plumbing and electrical studies, whether each company's website lets the AI assistants' crawlers read it.

The results, all 13 restoration companies

Sorted by prompts linked, then answers named. Mention counts in the names-instead column are across that company's twenty answers.

The table is available as a spreadsheet: download the study data as CSV, 13 rows, one per company. Every one of the 260 answers is published too; the data appendix at the end explains the file.

What the 260 answers say

1. One franchise was named two and a half times as often as all 13 local companies

SERVPRO was named in 142 of the 260 answers. The thirteen first-page local companies were named 56 times between them. In every one of the thirteen metros, SERVPRO appeared in at least five of the local company's twenty answers; in Denver it appeared in sixteen, in Charlotte fifteen, in Boston and Philadelphia fourteen. In seven metros it was the single most-named company in the local firm's own answer set.

It is the Roto-Rooter and Mr. Electric pattern at a different scale. Roto-Rooter was named in 73 of 260 plumbing answers; the two electrical franchises in 64 of 259. Restoration has five national names in the top six, and SERVPRO alone outscores every franchise from the earlier studies combined. A brand with a franchise page in every city, a location in Google's map pack, and a company website the assistants cited 213 times is the name an assistant reaches for when it knows little about the local market, and in restoration it knows very little.

2. Restoration lands between electricians and plumbers in the same cities

Same thirteen metros, same method, four trades. The totals: HVAC 80 of 260, electricians 67 of 259, restoration 56 of 260, plumbers 48 of 260. The restoration company was named more often than the plumber in the same city in six metros, less often in five, and the same in two.

Dallas is the standout: Dalworth Restoration, a large local company with decades of DFW history on its site, was named in fourteen of twenty answers, the best single result in any of the four October and November studies. Boston, Phoenix and Denver went the other way, zero in all three, and in all three the answer sets read like a franchise directory. Four trades in, the pattern holds: a first-page ranking is a fact about one trade in one city, and the assistants are answering from a different set of sources than Google ranks.

3. The most-cited source is the franchise's own website

In the three earlier trade studies the top source was an AI-era contractor directory: 196, 165 and 224 citations. In restoration the directories fall to the middle of the list and the top three sources are franchise websites: servpro.com with 213 citations, then the RestoPros and All Dry franchise sites. The tested companies' own sites were cited 54 times across all 260 answers. Lead marketplaces were cited 98 times, fewer than in any earlier study, because the assistants did not need them: the franchise sites already have a page for every city, written in the shape of an answer.

That is the mechanism in one line. An assistant asked about water damage in Denver finds a SERVPRO page for Denver, a PuroClean page for Denver and a 911 Restoration page for Denver before it finds anything a local company wrote, because the local company has one page about water damage and no page about Denver.

4. The companies named most had the most specific pages

The three companies linked in all five prompts, Dalworth, PureDry and Romexterra, have the deepest sites in the set: a page per damage type, a page per city or neighborhood, insurance and claims pages, and reviews with towns in them. The three companies named in none of twenty have single-page or near-single-page sites, two of which do not name their own metro on the homepage. The tool wrote national questions for those two because it could not tell where they work, and the assistants answered national questions with national brands.

5. Gemini and Claude name local companies; Perplexity least

The spread is narrower than in the earlier trades, and lower at the top: Gemini and Claude each named the tested company in 16 of 65 answers, against 23 of 65 for electricians and 31 of 65 for HVAC on Gemini. Perplexity, which cites sources the most heavily, named the first-page company in 9 of 65. Gemini has led or tied in all four same-city studies.

6. No site turned away an AI crawler

As in the plumbing and electrical studies, we asked each company's homepage for its page the way the assistants' crawlers do, from an ordinary office connection and again from a data center, and would have counted a block only when both agreed. Twelve of thirteen answered every crawler normally from both connections. The thirteenth, Ungerman, answered normally from the office connection, and the data-center check could not be completed before publication because our own checker rate-limited the request; it is marked inconclusive, not blocked. After two of thirteen plumbers and three of thirteen electricians, zero of thirteen is the best crawler result in the series, which makes the 56 of 260 harder to explain away: the assistants could read every one of these sites, and still named a franchise instead.

What this means for a restoration company

1. 01Name your market on the homepage, in text. Two of the three companies never named in an answer do not say where they work on their first screen, and the tool, like an assistant, wrote questions about the whole country. The metro, the counties and the towns, in words a crawler can read.
2. 02One page per damage type and per town. The franchise sites win because they have a Denver page, a water page, a fire page and a mold page each, in every city. A local company can hold the same ground in the towns it actually serves, with real jobs, real reviews and real drive times on each page. Restoration SEO is that list, and the water damage restoration SEO guide walks it in order.
3. 03Write the insurance and claims page. The prompts asked about burst pipes, floods and storm losses, and the answers lean on pages that explain deductibles, adjusters and documentation. Most local sites have none. Restoration marketing to insurance agents covers the agent side of the same work.
4. 04Be in the places the answers are built from. The Better Business Bureau, Expertise, Yelp and the restoration-specific directories in the chart, with a complete profile and the same name, address and phone everywhere. They sit behind the franchise sites in the citations, and they are open to any company.
5. 05Keep the crawlers in, and check it. None of these thirteen sites blocked an AI crawler, which is rare in our series. Keep it that way after the next security plugin update; the free AI crawler checker takes ten seconds.

To see where your company lands, run the AI visibility test on your own business, free, and we will send the twenty raw answers. The earlier studies: electricians, plumbing companies, HVAC companies and roofing companies. All five, with the numbers side by side, are on the research page.

Data appendix

Two files sit behind this article. The summary has 13 rows, one per company, including the crawler-access result. The answer-level file has 260 rows, one per answer, for all thirteen companies. Every number in the article can be recomputed from them.

How it was collected

• When: November 10, 2026, for the AI answers, the Google results and the crawler check.
• How: a third-party visibility-testing tool, called through its API. It writes the prompts from each company's website, sends each one to the four assistants with no account history and no location, and records what comes back.
• Models, as recorded: ChatGPT gpt-4o-mini, Perplexity sonar, Gemini gemini-3.1-flash-lite, Claude claude-haiku-5-5. These are the smaller API models, not necessarily what the consumer apps answer with.
• Crawler check: each homepage requested once per crawler, identifying as that crawler, from an office connection and again from a data-center connection through our own checker. A block counts only when both agree. One data-center check did not complete and is marked inconclusive.
• One run per company. Nothing was repeated, so the file is one day's answers, not an average.

What the columns mean

• own_site_linked: the tool's flag that the answer put the company's own website in front of the reader. A prompt counts toward linked, of 5, if any of its four answers has it.
• company_named: the company appears in the tool's list of brands named in that answer, link or not. This is named, of 20.
• cited_domains and cited_urls: the sources the tool recorded for the answer, the pages the assistant drew on. A source is not a click.
• answer_excerpt: the first 400 or so characters of the answer. The tool stores only that much.

Checks and edits

• Recomputed from the answer file, all thirteen companies match the published scores on both counts, and the total is 56 of 260 answers naming the tested company.
• The three lead marketplaces appear as [lead marketplace] in the sources, the links and the answer text, as in the earlier studies. Marketing-firm pages appear as [marketing firm, withheld]; we don't name other agencies. Nothing else is edited.
• The SERVPRO count of 142 is answers where the name appears in any form ("SERVPRO of Northwest Charlotte", "Servpro"); an answer naming two SERVPRO locations is counted once. The same rule gives Paul Davis 65, ServiceMaster 48, PuroClean 47 and 911 Restoration 46.
• Two companies tested and then excluded (one in Atlanta, one in Nashville) are not in either file; their raw results are kept in the data pack with the reason.

Both files are free to use under CC BY 4.0, with credit to Riptide Growth.