For home service and trades owners

Your next customer just asked ChatGPT for a plumber.

It answered with five names, and she called the first one. Is yours on that list? guidepost shows who's ahead, which sources the assistant trusted, and the first thing to fix.

Start weekly monitoring — $149/mo Request a free reading

Not the app on your phone. Results can differ from ChatGPT or Gemini on a phone. It never connects to your accounts or your software.

“Best HVAC company in Chino Hills”

  1. 1Valley Comfort Heating
  2. 2Foothill Air Services
  3. 3Ridgeline Heating & Air
  4. 4Summit Climate Co.
  5. Your name is not here yet
Example layout. Company names are illustrative.

In 2023 a homeowner with a burst pipe searched Google and called someone on page one. Today she asks her phone, gets three names, and calls the first one. Those three names come from somewhere: review sites, directories, and lists the assistant has learned to trust. guidepost finds out whether you're one of them. If you're not, it shows which of those sources would change the answer.

Three steps, once a week.

  1. 1

    It asks the real questions.

    The twelve things a homeowner in your city actually types, not keywords. “Best HVAC company in Chino Hills.” “My AC is blowing warm air, who should I call.” Each one asked three times on ChatGPT, Gemini and Perplexity, because answers vary between runs, and once on Claude.

  2. 2

    It reads the answers and traces the sources.

    ChatGPT, Claude, Gemini and Perplexity, recorded word for word. Whether you were named, where you ranked, who was named instead, and which pages the assistant was reading when it decided.

  3. 3

    It tells you what to fix first.

    The sources worth acting on, ranked, each showing the count it rests on. Not a score with no cause attached, and not a list of thirty tasks nobody will start.

From our own reading

In August 2026 we read 12 local markets around Chino Hills twice, a day apart: 151 home service companies, 432 answers recorded word for word each time. 18 of those companies carry 200 or more reviews. 6 of those 18 were named in none of the twelve questions in both readings, so it was not a bad morning. The most reviewed of them has 872.

A five-star record does not put you in the answer. What decides it is which pages the assistant read, and 61% of the time that is a home service company’s own website, usually a competitor’s. That part you can change. See all 12 markets, free.

What this does not do.

  • It doesn't connect to your accounts, your field service software, or your books.
  • It isn't the app on your phone. We measure OpenAI, Anthropic, Gemini and Perplexity with web search on. Those answers can differ from the consumer apps.
  • It can't make an assistant recommend you. Nobody controls what these models say. What it can do is show you which sources they're reading, which is the part you can change.
  • It won't promise a number. Answers move between runs, so every reading reports the spread instead of one confident figure.
  • If you're the only excavation contractor in a town of four thousand, the assistants already name you. Request a reading if you want one. Our founder looks at it himself, and nothing starts until he does, so keep your money.

Weekly evidence for $149 a month.

Twelve homeowner questions across OpenAI, Anthropic, Gemini and Perplexity. ChatGPT, Gemini and Perplexity are asked three times a week, Claude once. That is 120 recorded answers, and you see every one of them: who was named, which sources the model cited, and what changed since last week. Answers move on their own, so a single reading is a sample and the weekly repetition is what makes a change believable.

$149 per month, one location

Or $1,490 a year. Save $298. Same weekly readings.

  • 12 homeowner-style questions for your trade and city
  • OpenAI, Anthropic, Gemini and Perplexity with web search enabled, including Claude, which the free reading does not cover
  • 3 runs per question on ChatGPT, Gemini and Perplexity, 1 on Claude, 120 observations each week
  • Every raw answer, recommendation, competitor and cited source
  • One evidence-backed action and a weekly owner email

Start weekly monitoring — $149/mo Monthly, yearly, and the $499 job

Billed through Stripe. Cancel monthly in the portal, or pay once for the year. No account integrations.

Or start with one free reading.

Send your website and an email. City helps if you have it. We look at the request and, if a reading makes sense, our founder runs it by hand. You then get a one page report: every question asked, the answers word for word, the sources each assistant used, and the first thing to fix. It's a single reading across ChatGPT, Gemini and Perplexity, taken once per assistant, and it says so on the page. Weekly tracking adds Claude and asks each question three times. Free, and nothing to install.

Email and website. You'll hear from our founder. Nothing starts until he does. No spam. Your data is never used to train AI. Privacy

Nothing to install and no permissions to grant. Not the app on your phone. It never connects to your accounts or your software. One form tells us you want a reading. You'll hear from our founder. Nothing starts until he does.