Original research · 2026-07 edition

AI SEO Statistics: Solar (2026-07 edition)

40 questions · 120 AI responses · 3 models · measured 2026-07-06

See how AI talks about YOUR businessFull data table

The question bank

The questions we tested — sampled from real buyer journeys in solar.

Each model answered every question once, same wording, same day. These are the prompts behind every percentage on this page.

How do I know if my roof gets enough sun for solar panels to actually pay off?
Is it cheaper to buy solar panels outright or go with a lease program?
What is the average payback period for a residential solar system in a sunny state?
Can I still get solar if my roof is about 10 years old or should I replace the roof first?
What are the biggest red flags when talking to a solar door-to-door salesman?
How does net metering work and does every utility company offer it?
I have a 300 dollar monthly electric bill so how many panels would I realistically need to get that to zero?
What happens to my solar power during a blackout if I do not have a battery backup?
Show all 40 questions
Are there any federal tax credits left for this year and how do I actually claim them?
Is it worth getting a home battery storage system or is it just an expensive add-on?
What is the difference between monocrystalline and polycrystalline panels in terms of efficiency?
Can I install solar panels myself to save on labor costs or is that a bad idea for insurance?
Does adding solar panels really increase my property value or does it make the house harder to sell?
What kind of maintenance do solar panels need over a 20 year lifespan?
How do I compare two different solar quotes when the hardware and offset percentages are totally different?
My HOA has strict rules about how the front of the house looks so can they legally block me from installing solar?
Are there specific solar installers that specialize in flat roofs or tile roofs?
What is the deal with solar shingles versus traditional panels and are they actually durable?
If I move in five years what happens to my solar lease agreement?
How do I check if a solar contractor is actually licensed and insured in my specific state?
Will hail or heavy snow damage the panels and does my homeowners insurance cover that?
What is a PPA and why do some people say to avoid them at all costs?
Should I wait for better solar technology to come out next year or buy now?
How much square footage do I need on my roof for a 10kW system?
Are there any hidden costs like electrical panel upgrades that I should budget for?
What specific questions should I ask a solar company during the initial site visit?
Is ground-mounted solar significantly more expensive than roof-mounted solar?
Can I add more panels to my system later if I buy an electric car next year?
What is a realistic timeline from signing a solar contract to actually having the system turned on?
Do solar panels still produce any energy on cloudy or rainy days?
Are there any local rebates or state-specific incentives I should look for besides the federal ones?
What happens if the solar company I use goes out of business in a few years?
Is it better to finance through the solar company or get a home equity loan for the project?
How do I know if my inverter is sized correctly for my specific panel array?
Why is my neighbor's solar quote so much lower than the one I just got for the same size system?
Do I need to clean my solar panels regularly or does the rain take care of it?
Can I go completely off-grid with solar in a suburban area or is that illegal?
What is the difference between a string inverter and microinverters for a residential setup?
Are there any grants for low-income households to get solar installed for free?
How do I read a solar production estimate to make sure the company isn't overpromising on savings?

Model by model

22-point average divergence: which AI you ask changes the answer.

The divergence index is the average gap between the most and least likely model per behavior. Higher = the models disagree more about solar buyers.

Behavior rates across 40 solar buyer questions, 2026-07 edition. Last column: average across models.
ChatGPTClaudeGeminiConsensus
Recommends hiring a professional68%48%13%40%
Suggests DIY first23%18%13%80%
Names specific providers13%20%15%70%
Gives price or cost info18%35%48%53%
Tells to check reviews8%5%0%93%
Tells to verify credentials18%8%3%85%
Mentions case studies / portfolio5%8%0%90%
Mentions local proximity33%28%8%60%
Gives selection criteria35%25%18%60%
Warns about red flags8%15%15%88%
Asks a clarifying question60%50%0%20%
Recommends multiple quotes18%13%0%75%

By model

How each assistant handled Solar questions.

Reading the 120 answers model by model shows how differently the three assistants treat the same solar questions. On the most consequential behavior — whether to send the buyer to a professional at all — the rate ranged from 67.5% (ChatGPT) down to 12.5% (Gemini), a 55-point gap on an identical question set.

Across the 40 solar answers it produced, ChatGPT recommended hiring a professional in 67.5% of them and suggested a DIY approach first 22.5% of the time. It named a specific provider in 12.5% of answers (about 0.2 distinct providers per answer) and included price or cost information 17.5% of the time. ChatGPT asked a clarifying question before answering in 60% of cases, warned about red flags or scams in 7.5%, and told the buyer to verify credentials in 17.5%, averaging 549 words per answer. On the remaining cues it told the buyer to check reviews in 7.5%, pointed to case studies or a portfolio in 5%, and framed the choice around local proximity in 32.5%; a selection-criteria checklist appeared in 35% of its answers and a recommendation to gather multiple quotes in 17.5%.

Across the 40 solar answers it produced, Claude recommended hiring a professional in 47.5% of them and suggested a DIY approach first 17.5% of the time. It named a specific provider in 20% of answers (about 0.4 distinct providers per answer) and included price or cost information 35% of the time. Claude asked a clarifying question before answering in 50% of cases, warned about red flags or scams in 15%, and told the buyer to verify credentials in 7.5%, averaging 301 words per answer. On the remaining cues it told the buyer to check reviews in 5%, pointed to case studies or a portfolio in 7.5%, and framed the choice around local proximity in 27.5%; a selection-criteria checklist appeared in 25% of its answers and a recommendation to gather multiple quotes in 12.5%.

Across the 40 solar answers it produced, Gemini recommended hiring a professional in 12.5% of them and suggested a DIY approach first 12.5% of the time. It named a specific provider in 15% of answers (about 0.5 distinct providers per answer) and included price or cost information 47.5% of the time. Gemini asked a clarifying question before answering in 0% of cases, warned about red flags or scams in 15%, and told the buyer to verify credentials in 2.5%, averaging 264 words per answer. On the remaining cues it told the buyer to check reviews in 0%, pointed to case studies or a portfolio in 0%, and framed the choice around local proximity in 7.5%; a selection-criteria checklist appeared in 17.5% of its answers and a recommendation to gather multiple quotes in 0%.

Taken together, ChatGPT is the assistant most likely to route a solar buyer to a professional (67.5%) and Gemini the least (12.5%). ChatGPT produced the longest answers, at 549 words on average. Specific providers were named most often by Claude (20%) — even there, roughly one answer in 5 carried a name.

Where they disagree

The behaviors where the choice of model changes the answer.

The divergence index for this study is 21.5 points — the average distance between the most and least likely model across the coded behaviors. The gaps below are where which assistant a solar buyer happens to ask matters most:

  • Asks a clarifying question: from 0% (Gemini) to 60% (ChatGPT) — a 60-point spread.
  • Recommends hiring a professional: from 12.5% (Gemini) to 67.5% (ChatGPT) — a 55-point spread.
  • Gives price or cost information: from 17.5% (ChatGPT) to 47.5% (Gemini) — a 30-point spread.
  • Mentions local proximity: from 7.5% (Gemini) to 32.5% (ChatGPT) — a 25-point spread.
  • Gives selection criteria: from 17.5% (Gemini) to 35% (ChatGPT) — a 18-point spread.

The widest single gap — asks a clarifying question, 60 points — means a solar buyer can receive materially different guidance on the same question depending only on which assistant they happen to open, so any visibility strategy built on a single model's behavior describes only part of the solar market.

Where they agree

The points of near-consensus in Solar.

On other behaviors the three models move almost in lockstep — the points of near-consensus for solar, where all three landed within a few points of each other:

  • Names a specific provider: 12.5%–20% across all three (a 8-point spread).
  • Tells the buyer to check reviews: 0%–7.5% across all three (a 8-point spread).
  • Mentions case studies or portfolio: 0%–7.5% across all three (a 8-point spread).
  • Warns about red flags or scams: 7.5%–15% across all three (a 8-point spread).

Measured question by question, the three assistants coded a response the same way most consistently on "tells the buyer to check reviews" (identical coding in 92.5% of questions) and least consistently on "asks a clarifying question" (20%).

Every behavior, measured

All twelve coded behaviors for Solar, averaged across the three models.

The behaviors AI models reproduce most often for solar are recommends hiring a professional (42.5% on average), asks a clarifying question (36.7%) and gives price or cost information (33.3%); the rarest are mentions case studies or portfolio (4.2%), tells the buyer to check reviews (4.2%) and tells the buyer to verify credentials (9.2%). Each figure below is the share of a model's 40 answers in which the behavior appeared at least once, averaged across the 3 models with the full per-model range in parentheses:

  • Recommends hiring a professional: 42.5% on average (ChatGPT 67.5%, Claude 47.5%, Gemini 12.5%) — a 55-point spread.
  • Asks a clarifying question: 36.7% on average (ChatGPT 60%, Claude 50%, Gemini 0%) — a 60-point spread.
  • Gives price or cost information: 33.3% on average (ChatGPT 17.5%, Claude 35%, Gemini 47.5%) — a 30-point spread.
  • Gives selection criteria: 25.8% on average (ChatGPT 35%, Claude 25%, Gemini 17.5%) — a 18-point spread.
  • Mentions local proximity: 22.5% on average (ChatGPT 32.5%, Claude 27.5%, Gemini 7.5%) — a 25-point spread.
  • Suggests a DIY approach first: 17.5% on average (ChatGPT 22.5%, Claude 17.5%, Gemini 12.5%) — a 10-point spread.
  • Names a specific provider: 15.8% on average (ChatGPT 12.5%, Claude 20%, Gemini 15%) — a 8-point spread.
  • Warns about red flags or scams: 12.5% on average (ChatGPT 7.5%, Claude 15%, Gemini 15%) — a 8-point spread.
  • Recommends multiple quotes: 10% on average (ChatGPT 17.5%, Claude 12.5%, Gemini 0%) — a 18-point spread.
  • Tells the buyer to verify credentials: 9.2% on average (ChatGPT 17.5%, Claude 7.5%, Gemini 2.5%) — a 15-point spread.
  • Tells the buyer to check reviews: 4.2% on average (ChatGPT 7.5%, Claude 5%, Gemini 0%) — a 8-point spread.
  • Mentions case studies or portfolio: 4.2% on average (ChatGPT 5%, Claude 7.5%, Gemini 0%) — a 8-point spread.

Trust signals

How well the models protect the solar buyer.

Beyond whether to hire, the rubric codes how carefully each assistant protects the solar buyer once a decision is made. Telling the buyer to check reviews or ratings appeared in 4.2% of answers on average. Verifying credentials or certifications appeared in 9.2%. Warning about red flags or scams appeared in 12.5%.

On structuring the decision, a selection-criteria checklist showed up in 25.8% of answers on average and a recommendation to gather multiple quotes in 10%. The single least-reproduced protective signal for solar is "tells the buyer to check reviews" at 4.2% on average — the clearest opening for content that supplies it, since the models are not yet reliably surfacing that guidance on their own.

Referral behavior

Do AI models name Solar providers?

For service providers the decisive question is whether these systems name anyone at all. Across 120 solar answers, a specific provider was named in 15.8% of responses on average — roughly 0.4 distinct providers per answer. In practice the assistants behave far more as an explanatory layer than as a referral engine for solar: visibility comes from being the reasoning a model reproduces, not from being the named recommendation.

The question set

What these 40 Solar questions cover.

The 40 questions behind every percentage on this page were drawn from real solar (home services; buyer hiring decisions for this specific service) buyer journeys. Each was put to all 3 models once, with identical wording, so the rates above describe how the assistants handled this exact solar question set — not a general prior or a hand-picked subset. The full list is shown earlier on this page; the coded percentages are what those specific questions produced.

How to read this

A note on the numbers.

A percentage here is the share of a model's 40 answers in which the behavior appeared at least once — not a confidence score. Because each model answered every question exactly once on 2026-07-06, the figures describe this specific solar question set and snapshot rather than a general prior. The full protocol and coding rubric are documented in the study methodology.

Methodology

A controlled snapshot, documented end to end.

40 standardized buyer questions per industry, one response per model per question (ChatGPT (gpt-5-mini), Claude (claude-sonnet-5), Gemini (gemini-3-flash-preview)), collected 2026-07-06, coded against a fixed 12-behavior rubric with human QA. AI outputs vary with model version, location and time — figures describe this sample and window, and are refreshed each edition. Read the full methodology →