How Solar Installers Get Recommended by AI

By the alphaa team — we run AI-visibility scans across thousands of businesses, including home-services and solar contractors. Last updated 26 August 2026.

Short answer: AI engines recommend the solar installers whose licence and certification details, service area, and review history are stated consistently and in plain text across the open web — on your own site, on your Google Business Profile, on the state licence register, on EnergySage and similar marketplaces, and in local news or trade coverage. Solar is a high-consideration, high-fraud-anxiety purchase, so assistants lean unusually hard on verifiable trust signals and on third-party sources rather than on your marketing copy. If your NABCEP certification, your licence number, and the counties you actually cover are only visible inside a slide-out menu or a PDF, an AI engine cannot use them, and it names a competitor who spelled them out.

What a homeowner actually asks

The queries we see people bring to assistants about solar are longer and more suspicious than typical local search. They look like:

• "Who are the most reputable solar installers in Sacramento County?"
• "Is [company name] a legitimate solar company or a lead reseller?"
• "Which solar installers near me handle their own permitting instead of subcontracting?"
• "What should I ask a solar company before signing a 25-year agreement?"
• "Solar company that services rural properties outside city limits"

Notice how few of these are the phrase "solar panel installation near me." The customer is pre-screening for legitimacy, scope, and business model. That is the conversation you need to be present in — and the way to be present in it is to have answered those exact questions somewhere an engine can read.

Why solar is harder than other trades

We look at a lot of local-services scans, and solar behaves differently from plumbing or roofing in three specific ways:

1. The category has a reputation problem, and models know it. Years of aggressive door-knocking, misrepresented leases, and high-profile installer bankruptcies are all over the training data and the live web. Ask an assistant about solar and it will often volunteer warnings before it volunteers names. That means your content has to clear a higher trust bar than a roofer's does.
2. Aggregators dominate the visible surface. EnergySage, SolarReviews, marketplace comparison pages, and state programme sites are the pages that get cited when someone asks for installers in an area. Being absent from those is a bigger handicap than being weak on your own blog.
3. Incentives change constantly, and stale pages get discounted. Federal and state incentive terms shift, and pages that describe expired programmes as current are worse than useless — they signal that the site is not maintained. We wrote about why this matters generally in content freshness and AI search; in solar it is acute.

The seven things to fix, in order

1. Put your licence and certifications in readable text

This is the single most common miss. Contractor licence number, the state that issued it, NABCEP certification for your installers, and any manufacturer certifications (Tesla Certified Installer, Enphase Installer Network, and so on) should appear as text on your About and Contact pages — not baked into a badge image, not in a footer graphic, not in a downloadable capability deck. AI engines do not reliably extract text from images, and a licence number is exactly the kind of verifiable detail that makes a model comfortable naming you. See what AI engines can and cannot read in PDFs and images for the mechanics.

Write it plainly: "Licensed electrical contractor, [State] licence #XXXXXX. Three NABCEP-certified PV Installation Professionals on staff." That sentence is liftable. A badge is not.

2. State your service area as a list of named places

"We serve the greater metro area and surrounding communities" tells a model nothing. A model answering "solar installers that cover Placer County" needs to see the string "Placer County." List the counties and the towns you genuinely work in, and say plainly where you stop — including any distance limit or minimum system size for outlying areas. Being explicit about what you do not cover is a trust signal, and it prevents the worse outcome of being recommended to someone you cannot serve.

3. Answer the business-model question before you are asked

Homeowners want to know: do you do the work, or do you sell the lead? Do your own crews do the install, or do you subcontract? Do you handle permitting, interconnection paperwork, and utility approval in-house? Is the sale a cash purchase, a loan, a lease, or a PPA — and which of those do you actually offer? Put this on a page in the words a customer would use. Very few installers do, which means the ones who do own that entire class of question.

4. Get your marketplace and directory presence consistent

Claim and complete your EnergySage and SolarReviews profiles, your Google Business Profile, and your state's licensed-contractor listing. Make sure the business name, address, phone number, and service area match your website character for character. Inconsistent details across sources are the fastest way to get an entity fragmented, so the engine is no longer confident which record refers to you. That mechanism is covered in directory listings and NAP citations.

5. Build the local-conditions content nobody else writes

Generic "benefits of solar" posts are the most duplicated content on the internet and get cited almost never. What gets cited is specific and local: how your utility's current net-metering successor tariff changes payback maths, what your county's permitting timeline actually looks like in practice, how HOA rules interact with the state solar-rights statute where you operate, what a snow-load or wildfire-zone requirement means for mounting in your area. You know these things because you do the work. That first-hand specificity is precisely what content that AI engines quote looks like.

6. Publish real project details, with the boring numbers

A case study that says "happy customer in Fresno" is not evidence. One that says "7.2 kW system, 18 panels, south-west facing composition-shingle roof, permitted in 11 days, first full-year production 10,900 kWh against a modelled 11,200" is. Include the case where production came in under model and explain why. Honest numbers, including the disappointing ones, are the strongest trust signal a solar company can publish, and they give an assistant something concrete to summarise.

7. Fix your reviews as a body of text, not a score

Engines read review content, not just the star average. Reviews that mention your permitting handling, your responsiveness after install, and your warranty service teach a model what you are good at — which is how you get recommended for a specific need rather than for "solar." Ask satisfied customers to describe what actually happened rather than to leave five stars. More on the mechanism in why your Google reviews decide your AI visibility.

What to do about the trust problem, specifically

Because assistants often answer solar questions with caution first, the installers who win are the ones whose own content matches that cautious register. Concretely, that means publishing the things a sceptical buyer wants and most competitors avoid:

• A plain-language explanation of what your warranty does and does not cover, and who honours it if you cease trading.
• Your actual cancellation terms and any deposit policy.
• Whether your production estimate is a guarantee or a model, and what happens if actual output falls short.
• What happens to the agreement if the homeowner sells the house.
• A statement of who owns the incentives and tax credits under each financing option you offer.

This is the same posture we take about our own category: you cannot promise an outcome you do not control. AEO shapes the public signals AI engines read about you; it does not buy a placement, and anyone selling a guaranteed spot in ChatGPT is describing something that does not exist. Solar buyers have been burned by exactly that kind of promise, so matching their scepticism is a competitive advantage.

A 30-day sequence that fits around real work

1. Week 1 — audit. Ask ChatGPT, Gemini, Claude, and Perplexity the five questions at the top of this article with your county named. Record who gets named and which sources get cited. That list of cited sources is your actual target list.
2. Week 2 — fix the facts. Licence number, certifications, named service area, and business-model page in plain text on your site. Reconcile name, address, and phone across Google Business Profile, EnergySage, SolarReviews, and the state register.
3. Week 3 — publish two local pieces. One on your utility's current net-metering terms and what they mean for payback; one on your county's permitting and interconnection process with real timelines. Date them and commit to updating when the rules change.
4. Week 4 — evidence. Two project write-ups with real system sizes and production numbers, and a review request to your last ten customers asking them to describe the process, not to rate it.
5. Re-check at 60 and 90 days. Movement on this is measured in weeks to months, not days — see how long AEO takes to work.

Questions we get asked

Do I need to be on EnergySage to be recommended?

You do not need it, but it helps disproportionately in this vertical, because marketplace and review-site pages are heavily represented among the sources engines cite for "best solar installers in [place]." If you are philosophically opposed to lead marketplaces, at minimum make sure you are listed and accurate on the free review platforms and your state register.

Will a page targeting every town I serve work?

Only if each page contains something genuinely specific to that town — the local permitting office, the utility, actual projects there. Thin duplicated city pages are a well-known pattern and add little. We go through the evidence in do city landing pages work for AI search.

My reviews mention a bad install from two years ago. Is that fatal?

No, and trying to bury it usually backfires. A public, specific response describing what you changed is read as evidence of accountability. Volume and recency of good detailed reviews matter more than the absence of a bad one.

Does it matter that incentive rules keep changing?

It matters a great deal, and it is an opportunity. Most installer sites still describe superseded programmes. A page that is visibly current, with a stated last-updated date and the correct present terms, becomes the reliable source in a category full of stale ones.

Can I just run ads instead?

Ads and AI recommendations are separate systems — paid spend does not buy you a mention in an organic AI answer. See do paid ads affect AI recommendations.

The bottom line

Solar is a trust purchase, and AI assistants have absorbed the industry's trust problems along with everything else. You do not overcome that with better adjectives. You overcome it with verifiable, plainly stated facts — licence, certifications, named service area, honest financing terms, real project numbers, and current local rules — repeated consistently everywhere an engine looks. The installers getting named today are rarely the biggest ones. They are the ones who wrote the specifics down.

Run a free AI visibility scan →