We audited the homepages of 335 Google Maps solar installers in 26 US cities and compared the ones ChatGPT, Perplexity, Gemini, Claude and Google AI Mode recommended with the ones they skipped. Here are the 12 biggest differences, and the popular features that made none.
When we audited the homepages of 335 Google Maps solar installers across 26 US cities on 8 October 2026, 62.7% of the installers that at least one of five AI assistants recommended linked to a battery or storage page, compared with 33.9% of the installers the AIs skipped. Battery content produced the two largest gaps among the 26 homepage signals we measured, followed by showing reviews, stating years in business and adding business schema. Prices on the homepage, savings calculators, warranty lengths and FAQ schema made little or no difference.
TL;DR
- In our October 2026 audit of 335 solar installer websites in 26 US cities, 62.7% of the 102 installers named by ChatGPT, Perplexity, Gemini, Claude or Google AI Mode linked to a battery or storage page, against 33.9% of the 233 installers they did not name.
- The battery gap held for small installers too: among installers with fewer than 50 Google reviews, 66.7% of the AI-named ones had a battery page, against 30.6% of the rest.
- AI-named installers were also more likely to show reviews or testimonials (70.6% vs 48.9%), state their years in business (41.2% vs 22.7%) and carry business schema (75.5% vs 60.5%).
- Prices on the homepage (23.5% vs 21.5%), savings calculators (14.7% vs 13.7%), FAQ schema (9.8% vs 10.3%) and warranty lengths (18.6% vs 21.5%) differed by three points or less.
- AI-named installers had a median of 134 Google reviews against 26 for the rest, so these website signals travel with bigger, busier companies, and every gap here is a correlation from one snapshot.
This is Part 3 of the Orkkid Solar AI Study. In Part 1 we asked five AI assistants to recommend a solar installer in 26 US cities and matched every answer against the Google Maps top 20 for "solar installer" in that city. Here we take every Maps installer with a working website and ask a practical question: what is on the homepages of the installers the AIs picked that is missing from the homepages they skipped? The parent study for the whole method is our AI recommendation study. Orkkid sells SEO and GEO services to solar companies, so we have a stake in this answer, which is why every number below comes from data described in the method section.
The 12 signals at a glance
Each item below compares two groups of installer websites: the 102 that at least one AI named in its city (ChatGPT, Perplexity, Gemini, Claude or Google AI Mode) and the 233 that none of them named. The gap is the difference in percentage points, and the list runs from the largest gap down.
| # | Signal on the homepage | Named by AI (102) | Not named (233) | Gap (points) |
|---|---|---|---|---|
| 1 | Mentions batteries or backup power | 80.4% | 51.5% | 28.9 |
| 2 | Links to a battery or storage page | 62.7% | 33.9% | 28.8 |
| 3 | Shows reviews or testimonials | 70.6% | 48.9% | 21.7 |
| 4 | States years in business | 41.2% | 22.7% | 18.5 |
| 5 | Has a tap-to-call phone link | 82.4% | 63.9% | 18.5 |
| 6 | Business schema (LocalBusiness or similar) | 75.5% | 60.5% | 15.0 |
| 7 | Mentions NABCEP certification | 20.6% | 6.0% | 14.6 |
| 8 | Mentions financing, loans or leases | 69.6% | 55.4% | 14.2 |
| 9 | Shows a license number | 28.4% | 15.0% | 13.4 |
| 10 | City name in the homepage title | 30.4% | 18.9% | 11.5 |
| 11 | Mentions the tax credit or incentives | 59.8% | 49.8% | 10.0 |
| 12 | Has an llms.txt file | 35.3% | 25.3% | 10.0 |
What AI-recommended solar installers put on their websites
Share of installer homepages with each signal: named by AI vs not named
The biggest gap: mentions batteries or backup power (80.4% of named installers vs 51.5% of the rest).
Source: Orkkid Solar AI Study, October 2026 (335 reachable installer websites)
1. Talk about batteries and backup power on the homepage
80.4% of AI-named installers vs 51.5% of the rest. Four in five AI-named installers mentioned batteries, home backup or a battery brand in their homepage text, and this was the largest gap in the audit. It fits what the AI answers themselves talked about: of the 130 AI answers in the study, 121 (93.1%) mentioned batteries or the Powerwall. Claude told a Los Angeles homeowner to "Ask about battery storage, since it matters more in California now that net metering has changed." Perplexity described one San Diego pick with "Consider it if backup power and battery sizing matter as much as panel production."
Homeowners scanning a homepage want to know quickly whether you do storage, and AI tools that summarize your site pick up whatever services the homepage states plainly.
What to put on the site: one homepage section that says, in a full sentence, that you design and install battery backup, with the brands you install named and a link to the battery page in item 2.
2. Give batteries their own page
62.7% vs 33.9%. AI-named installers were nearly twice as likely to link from the homepage to a dedicated battery or storage page, a gap of 28.8 points, just behind the plain mention. For a homeowner, a battery page answers the question they are already asking: can this company keep my lights on and store what my panels make? For an AI assistant, a page at a clear address such as /solar-batteries gives it a specific, quotable fact to attach to your name when someone asks for battery expertise.
What to put on the site: a dedicated battery page linked from the homepage navigation, naming the batteries you install (for example Tesla Powerwall, Enphase IQ or Franklin), what backup covers, typical system sizes, how batteries work with your local utility's export rules, and photos of real installs.
3. Show reviews or testimonials where people can read them
70.6% vs 48.9%. Seven in ten AI-named installers put reviews or testimonials on the homepage. Reviews matter twice here. The installers the AIs named had far more Google reviews (more on that below), and the AI answers lean on review sources: in our Part 1 findings, EnergySage appeared in most answers, and the answers regularly quoted star ratings and review counts.
What to put on the site: three to six recent, named testimonials on the homepage that mention the city and the job (for example "12 kW system with two batteries in Mesa"), plus a link to your full Google reviews. Our guide on how many Google reviews a solar company needs has city-by-city benchmarks.
4. Say how long you have been in business
41.2% vs 22.7%. AI-named installers were almost twice as likely to state their age on the homepage, with phrases such as "since 2008" or "15 years of experience". The AI answers use this detail constantly: 100 of the 130 answers (76.9%) included a founding year or a similar line. Claude described one Sacramento option with "This could suit you if you want a long-established company with battery experience."
What to put on the site: one plain sentence near the top, such as "Installing solar in Denver since 2009, with 2,400 homes completed." Use your real founding year and a real install count.
5. Make the phone number tappable
82.4% vs 63.9%. This is one of three extra fields the audit recorded beyond the 23 in the summary table, and it tied with years in business on the size of the gap. A tap-to-call link (a tel: link) does nothing for an AI model on its own. We read it as a marker of a site that has been built for phones and kept up to date, which tends to go with companies that invest in marketing generally.
What to put on the site: your phone number in the header as a tap-to-call link on every page, matching the number on your Google Business Profile exactly.
6. Add business schema to the homepage
75.5% vs 60.5%. Most installer homepages carried some structured data describing the business (we counted LocalBusiness, HomeAndConstructionBusiness, Electrician, GeneralContractor, RoofingContractor, ProfessionalService, Organization, Corporation or Store types), and AI-named installers had it 15 points more often. Google explains how it reads this markup in its LocalBusiness structured data guide.
What to put on the site: LocalBusiness schema with your legal name, address, phone, service area, opening hours and the same details as your Google Business Profile. Our free LocalBusiness schema generator builds the code for you.
7. Name your NABCEP certification
20.6% vs 6.0%. Only 35 of the 335 sites mentioned NABCEP at all (21 named, 14 not named), so this is a small group, but AI-named installers were more than three times as likely to mention it. 47 of the 130 AI answers (36.2%) brought up NABCEP, usually as advice: Claude in Dallas said to "Check for NABCEP certification, the warranty terms (panels, inverter, and labor), and recent Google or BBB reviews."
What to put on the site: if your installers hold a NABCEP certification, say so on the homepage and name how many certified professionals you employ. Never claim a credential you do not hold.
8. Explain financing, loans and leases
69.6% vs 55.4%. Most installers mention financing somewhere, and AI-named installers did so more often. A dedicated financing page showed a much smaller gap (32.4% vs 27.5%), so in our data the plain mention on the homepage mattered more than a separate page.
What to put on the site: a short homepage section listing the ways a homeowner can pay (cash, loan, lease or PPA, where you offer them) in plain words, linked to a page that explains the trade-offs honestly.
Get your free GEO audit for solar companies
9. Print your contractor license number
28.4% vs 15.0%. AI assistants often tell homeowners to verify the license. 70 of the 130 answers (53.8%) mentioned licensing, and Claude told a Los Angeles homeowner to "Verify the contractor's license on the California CSLB website." An installer whose license number is on the homepage makes that check easy for the homeowner and gives the AI a verifiable credential.
What to put on the site: your state license number (for example a CSLB, ROC or electrical contractor number) in the footer of every page and on the about page, with a link to the state lookup.
10. Put your city in the homepage title
30.4% vs 18.9%. The homepage title tag is the line that shows in Google results and browser tabs. AI-named installers were more likely to include the city they serve in it. Search engines and AI tools both use the title to understand what a page is about and where the business works.
What to put on the site: a title such as "Solar Panel and Battery Installer in Tampa, FL | Company Name". If you serve several metros, give each its own location page with its own title.
11. Keep incentive information current
59.8% vs 49.8%. AI-named installers were more likely to mention the tax credit, incentives or rebates. There is a catch every solar marketer should know: according to the IRS, the 30% Residential Clean Energy Credit is not available for property placed in service after December 31, 2025. Our audit counted any mention, so some of these homepages may be showing outdated incentive claims.
What to put on the site: a dated incentives page that covers what is available in your state and utility today, updated whenever the rules change. AI tools and homeowners both trust a page that says when it was last checked.
12. Publish an llms.txt file
35.3% vs 25.3%. An llms.txt file is a plain-text summary of your site at yoursite.com/llms.txt, written for AI tools. A third of AI-named installers had one, against a quarter of the rest. No AI company has confirmed that it reads the file, so treat it as a cheap, low-risk addition and keep your expectations modest.
What to put on the site: a short llms.txt that lists your services, service area, license number, years in business and links to your battery, financing and reviews pages. Our llms.txt generator writes a first draft for you.
The more of these signals, the more often AI named the installer
We counted how many of the 12 signals each homepage had. Sites with seven or more were named by at least one AI 55.1% of the time (54 of 98), against 14.3% (13 of 91) for sites with three or fewer. Among installers with fewer than 50 Google reviews, 12 of the 36 sites with seven or more signals were named (33.3%), against 6 of 66 with three or fewer (9.1%).
Installers with more of the 12 signals were named by AI more often
Share of installer websites named by at least one of five AIs, by how many of the 12 listed signals the homepage had
55.1% of sites with seven or more of the 12 signals were named by AI, against 14.3% of sites with three or fewer.
- 0 to 3 of the 12 signals14.3%
13 of 91 sites named
- 4 to 6 of the 12 signals24%
35 of 146 sites named
- 7 or more of the 12 signals55.1%
54 of 98 sites named
Source: Orkkid Solar AI Study, October 2026 (n = 335 reachable installer websites in 26 US cities, 102 named by AI)
Bigger companies or better websites? What the data can and cannot say
The installers the AIs named were different businesses as well as different websites. The table below compares the two groups on the things a website cannot change overnight.
| Measure | Named by AI (102 sites) | Not named (233 sites) |
|---|---|---|
| Median Google reviews | 134 | 26 |
| Average Google star rating (rated listings) | 4.79 | 4.68 |
| Median position in the Google Maps top 20 | 8 | 11 |
| Median words on the homepage | 1,152.5 | 949 |
| Median years in business (sites that state it) | 18 (41 sites) | 16.5 (52 sites) |
| Links to a battery or storage page | 62.7% | 33.9% |
A median of 134 reviews against 26 is a big difference, and it matters for reading every gap in this post. Sites with a battery page had a median of 72 Google reviews, against 34.5 for sites without one. Sites with seven or more of the 12 signals had a median of 75 reviews, against 19 for sites with three or fewer. Larger, longer-established installers tend to have more reviews, more services and more developed websites all at once, and AI assistants may be responding to the size of the company as a whole.
So we split the battery-page gap by review count. It held in every band: among installers with fewer than 50 reviews, 66.7% of the AI-named ones had a battery page against 30.6% of the rest, and the pattern repeated at 50 to 199 reviews (60.5% vs 35.2%) and at 200 or more (62.5% vs 45.7%). The under-50 group has only 24 named installers, so read that band as a direction of travel with a wide margin of error. The gap also held city by city: in 20 of the 24 cities where both groups existed, the named installers were more likely to have a battery page.
Does the battery-page gap hold for smaller installers?
Share of homepages linking to a battery or storage page, AI-named vs not named, by Google review count
AI-named installers linked to a battery page more often in every review band, including installers with under 50 reviews (66.7% vs 30.6%).
Source: Orkkid Solar AI Study, October 2026 (n = 335 reachable installer websites in 26 US cities, 102 named by AI)
What this cannot show is cause and effect. This is one snapshot of one homepage per installer, compared with one run of each AI per city. A battery page may help an AI match your company to battery questions, or it may simply be one more sign of a company that offers more and markets more. The practical reading is the same either way: the installers AI recommends tend to state their services, credentials and track record plainly, and battery content is the clearest example. For how AI tools read and cite installer sites in the first place, see the sites AI reads when it recommends solar installers.
Smaller gaps worth knowing
These signals leaned toward AI-named installers too, by less than ten points each.
- Mentions Google reviews or a Google rating: 24.5% vs 14.6%.
- Has a blog or learning center: 61.8% vs 53.6%. Educational pages also earn citations in Google's AI Overviews, as our study of 60 solar searches showed.
- Mentions EnergySage: 6.9% vs 0.4%, which is only 7 sites against 1, so it is too small to read much into.
- Offers a free quote or assessment: 50.0% vs 43.8%.
- Has a chat widget: 19.6% vs 13.7%.
- Review stars in schema: 13.7% vs 8.2%.
- Has a financing page: 32.4% vs 27.5%.
- Lists service areas: 34.3% vs 30.0%.
What made no difference in our data
Several features that solar marketers spend real money on showed a gap of two points or less, or a negative one, between installers the AIs named and those they skipped.
| Signal on the homepage | Named by AI | Not named | Gap (points) |
|---|---|---|---|
| Shows prices on the homepage | 23.5% | 21.5% | 2.0 |
| Blocks an AI crawler in robots.txt | 2.9% | 1.7% | 1.2 |
| Has a savings calculator | 14.7% | 13.7% | 1.0 |
| FAQ schema | 9.8% | 10.3% | -0.5 |
| Mentions the BBB | 8.8% | 9.4% | -0.6 |
| States a warranty length | 18.6% | 21.5% | -2.9 |
Warranty is the most surprising one, because the AI answers talk about warranties all the time: 113 of the 130 answers (86.9%) mentioned them. Stating a warranty length on the homepage did not separate the two groups, and the median warranty stated was 25 years in both. When every installer quotes the same number, it stops being a reason to pick one of them. FAQ schema is similar: it appeared on 21.8% of Google Maps top-three listings, so it may relate to Maps ranking, but it did not relate to AI recommendations in our sample. Very few sites blocked an AI crawler (2.1% overall), too few to judge.
None of this means you should remove these features. A savings calculator, upfront pricing and a clear warranty can still turn visitors into booked consultations, and homeowners appreciate all three. Our data only says they did not distinguish the installers AI recommended. You can see all 23 summary signals, including how often each appeared among Maps top-three listings, in the table below.
23 website signals across solar installer homepages
Sort by the gap between installers AI named and those it skipped
| Mentions batteries or backup power | 60.3% | 78.2% | 80.4% | 51.5% | 28.9 |
| Has a battery or storage page | 42.7% | 50.9% | 62.7% | 33.9% | 28.8 |
| Shows reviews or testimonials | 55.5% | 61.8% | 70.6% | 48.9% | 21.7 |
| States years in business | 28.4% | 27.3% | 41.2% | 22.7% | 18.5 |
| LocalBusiness schema | 65.1% | 69.1% | 75.5% | 60.5% | 15.0 |
| Mentions NABCEP certification | 10.4% | 14.5% | 20.6% | 6% | 14.6 |
| Mentions financing, loans or leases | 59.7% | 65.5% | 69.6% | 55.4% | 14.2 |
| Shows a license number | 19.1% | 16.4% | 28.4% | 15% | 13.4 |
| City name in homepage title | 22.4% | 25.5% | 30.4% | 18.9% | 11.5 |
| Mentions the tax credit or incentives | 52.8% | 63.6% | 59.8% | 49.8% | 10.0 |
| Has an llms.txt file | 28.4% | 36.4% | 35.3% | 25.3% | 10.0 |
| Has a blog or learning center | 56.1% | 56.4% | 61.8% | 53.6% | 8.2 |
| Mentions EnergySage | 2.4% | 3.6% | 6.9% | 0.4% | 6.5 |
| Offers a free quote or assessment | 45.7% | 49.1% | 50% | 43.8% | 6.2 |
| Review stars in schema | 9.9% | 9.1% | 13.7% | 8.2% | 5.5 |
| Has a financing page | 29% | 25.5% | 32.4% | 27.5% | 4.9 |
| Lists service areas | 31.3% | 30.9% | 34.3% | 30% | 4.3 |
| Shows prices on the homepage | 22.1% | 27.3% | 23.5% | 21.5% | 2.0 |
| Blocks an AI crawler | 2.1% | 3.6% | 2.9% | 1.7% | 1.2 |
| Has a savings calculator | 14% | 18.2% | 14.7% | 13.7% | 1.0 |
| FAQ schema | 10.1% | 21.8% | 9.8% | 10.3% | -0.5 |
| Mentions the BBB | 9.3% | 10.9% | 8.8% | 9.4% | -0.6 |
| Mentions a warranty length | 20.6% | 18.2% | 18.6% | 21.5% | -2.9 |
Source: Orkkid Solar AI Study, October 2026 (335 reachable websites)
What to do this month
If you run marketing for a solar company, here is the short version, ordered by the size of the gap:
- Describe your battery and backup work on the homepage, give it its own page, and name the batteries you install.
- Put recent reviews, your years in business, your license number and any NABCEP certification on the homepage in plain sentences.
- Add LocalBusiness schema that matches your Google Business Profile, put your city in the homepage title and keep your incentive copy dated and current.
- Keep growing your Google reviews, because review count was the biggest difference between the two groups of all.
Then check whether AI assistants mention you by asking them yourself, city by city, the way we did. Our solar company SEO guide covers the rest of the site, and our generative engine optimization page explains how we approach AI recommendations for clients. If you want this done for your company, see our work with solar companies.
How we got this data
- Dates: AI answers and Google Maps results were collected on 7 and 8 October 2026; the website audit ran on 8 October 2026.
- AI engines and models: ChatGPT (gpt-5.5), Perplexity (sonar-pro), Gemini (gemini-3.8-flash) and Claude (claude-sonnet-5-5), each with web search on, through DataForSEO's AI optimization API, plus Google AI Mode through DataForSEO's SERP API.
- Prompt: "Who is the best solar installer in {City}, {State}? Give me your top picks." One run per engine and city: 130 answers in 26 cities (Denver, Austin, Phoenix, Charlotte, Seattle, Tampa, Los Angeles, San Diego, Sacramento, Las Vegas, Houston, Dallas, San Antonio, Orlando, Miami, Atlanta, Raleigh, Salt Lake City, Albuquerque, Tucson, Boston, Newark, Baltimore, Chicago, Minneapolis and Honolulu).
- Installer list: the Google Maps top 20 for "solar installer" in each city (520 listings), of which 436 listed a website. An installer counts as "named by AI" when at least one of the five engines recommended it in that city's answer, matched by business name.
- Website audit: our script fetched each site's homepage, robots.txt and llms.txt (15-second timeout, one retry on server errors or timeouts). 335 homepages loaded (316 unique domains, since a few companies rank in more than one city); 101 could not be read because they blocked our request, timed out or returned an error. Those 335 sites are the sample: 102 named, 233 not named.
- How signals were detected: every signal comes from the homepage HTML only, using keyword patterns on the visible text or on internal links. A "battery page" means an internal link (same domain or relative, with image, PDF, CSS and script files excluded) whose address contains words such as battery, powerwall, energy storage, home storage, home backup or backup power. "Mentions batteries" means the visible text contains battery, batteries, Powerwall, Enphase IQ, Franklin or home backup. Reviews means the word review or testimonial; years in business means a phrase such as "15 years of experience" or "since 2009"; license number means patterns such as "license #", "CSLB" or "ROC #"; prices means "per watt" or a dollar amount in the thousands; warranty length means a two-digit number of years within 40 characters of "warranty"; schema signals come from the JSON-LD on the homepage; tap-to-call means a
tel:link; llms.txt means a plain-text file at /llms.txt. - Limits: keyword matching can miss signals that sit on inner pages or load only with JavaScript, and it can occasionally count an unrelated word (for example a street called Franklin). Groups are small in places (for example 24 named installers with under 50 reviews). Each AI was asked once per city, so this is a snapshot, and every gap is a correlation. The full study method, including counting rules, is in our AI recommendation study.
Find out if AI recommends your solar company
One clear report on who AI recommends in your market right now, what earns them the spot, and the fastest path to getting your solar company named, delivered within 72 hours.
Tested across ChatGPT · Perplexity · Google AI Overviews · Gemini
Get my free GEO auditFAQ
What do AI-recommended solar installers have on their websites?
In our October 2026 audit of 335 solar installer homepages, the installers named by ChatGPT, Perplexity, Gemini, Claude or Google AI Mode were most different from the rest in mentioning batteries (80.4% vs 51.5%), linking to a battery or storage page (62.7% vs 33.9%), showing reviews (70.6% vs 48.9%), stating years in business (41.2% vs 22.7%) and linking a tappable phone number (82.4% vs 63.9%).
Will adding a battery page get my solar company recommended by ChatGPT?
Nobody can promise that. The battery-page gap held across review bands and in 20 of 24 cities, but AI-named installers also had far more Google reviews (median 134 vs 26). A battery page is a sensible, low-cost step that answers a question 93.1% of the AI answers raised, and it works best alongside a growing review count.
Do savings calculators or homepage prices help with AI recommendations?
Not in our data. Savings calculators appeared on 14.7% of AI-named installer sites and 13.7% of the rest, and homepage prices on 23.5% and 21.5%. They can still help homeowners decide and convert, so keep them if they work for your leads.
Does schema markup help solar companies appear in AI answers?
LocalBusiness and similar business schema showed a 15-point gap (75.5% of named installers vs 60.5% of the rest), and you can build it with our schema generator. FAQ schema showed no gap (9.8% vs 10.3%), although it was more common among Google Maps top-three listings (21.8%).
How was this solar website audit done?
We fetched the homepage of every Google Maps top-20 solar installer with a website in 26 US cities on 8 October 2026 and detected 26 signals with keyword patterns on the homepage text and internal links. We then compared the 102 installers that at least one of five AI assistants recommended with the 233 that none recommended.
Is this study still current?
The data is from October 2026. We plan to re-run the same prompts, cities and audit every quarter and update this post with the new numbers, so the gaps can be tracked over time.
