AI Insights Home Services & Construction
Fred: A Booking Assistant That Quotes Nothing It Can’t Honor
Talk to Fred
Ask Fred about Home Services & Construction
This is the same Fred you would put on your own site. Ask about Home Services & Construction, compliance, or how the guardrails work. Fred listens.
A homeowner lands on your site at night, describes a leak under the kitchen sink, and asks what the repair will run. A general-purpose chatbot answers with a number, "probably around $400 to $600," and books the slot. The crew shows up, finds rotted subfloor and a failed shutoff, and now the customer is waving a screenshot of the chat insisting on the price the website promised. The business wanted a tool to capture after-hours leads. It got one handing out binding quotes on jobs nobody had seen.
The threat and standard articles in this series walk through how that turns into a dispute. This one is the answer: what a compliant AI assistant for home services looks like, and why the design, not a disclaimer, is what keeps a website estimate from becoming a promise the crew has to eat.
The Real Choice Is Governed or Ungoverned
Contractors ask whether to use AI on the site at all. Homeowners already decided, they expect to book, ask about service areas, and get a sense of next steps without waiting until business hours, and the company that makes them wait loses the job. The question that matters is whether the assistant is governed.
An ungoverned chatbot blurts out a price, commits to a timeline, and texts the customer back in ways nobody cleared. A governed one books the visit, answers the logistics, and routes every price and scope question to the person who will actually do the work. Same convenience, completely different exposure when the invoice and the website disagree.
How Fred Holds the Line
Fred starts from the opposite default of a general chatbot. A generic bot answers anything unless told not to; Fred answers only what it is cleared to, and the boundary lives in the system rather than a prompt that drifts when a customer pushes.
In practice, Fred works from your own content. It books and reschedules, explains service areas and what you do, sets expectations about how an estimate works, and captures the lead with the details the crew needs. It does not hand out a price on a job it cannot see, because a number stated on your site can become an express warranty under UCC 2-313, an expectation the customer is entitled to hold you to. And it does not fire off marketing texts or calls on its own, since unsolicited automated contact runs straight into the TCPA’s consent rules. Fred runs more than 50 industry guardrail packs, and the home-services pack is built around the line between booking a job and quoting one.
The design proves itself under pressure. A homeowner who wants a number will rephrase "how much" several ways, and a prompt-instructed bot eventually answers the version it was not warned about. Fred does not rely on catching the phrasing. It decides what may be said before the reply forms, so the invented quote never reaches the customer.
Fred vs. a Generic AI Chatbot
| Situation | Generic AI Chatbot | Fred |
|---|---|---|
| "How much to fix this?" | Quotes a number nobody validated | Books the visit; leaves the estimate to the crew |
| "Can you guarantee it’s done Friday?" | Commits to a timeline | Captures the request; lets the office confirm |
| Following up by text | May blast automated texts, risking TCPA | Follows your consent rules; no rogue outreach |
| A binding promise | Creates one under UCC 2-313 | Sets expectations without committing a price |
| Where the rules live | In a prompt the customer can push past | Built into the system; enforced before output |
| Who owns the regulated answer | Effectively the chatbot, and your business | The crew that can actually see the job |
That single screen is the argument. A general chatbot is helpful until helpful becomes a binding quote or an unwanted text. Fred is helpful across everything that moves a job forward and structurally incapable of promising what the crew has not seen.
What Your Business Actually Gets
Set the rules aside and look at the phone. Fred handles the after-hours questions that go to voicemail and die there, do you cover my area, what do you charge to come out, can someone come Tuesday, what should I have ready, and it answers instantly, overnight, in plain language. It captures the lead with the context that makes the first call productive, so the estimator is not starting from zero. Everything that commits a price or a date stays with a human who can stand behind it.
The captured-lead math is the real win. The job that comes in at 9 p.m. and gets a useful, accurate response is the job that books, instead of the one that called the next contractor while waiting for a callback. Fred turns the after-hours window from a black hole into a pipeline, without ever putting a number on the table it cannot defend.
So the question is not whether your competitors will put AI on their sites. They will. It is whether yours is governed before a homeowner holds it to a price it never should have given.
Frequently asked questions
Can a compliant assistant give any kind of estimate?
It can explain how your estimates work, what a service call generally involves, and what factors affect a job, the same context a good intake person gives, and it routes the actual number to someone who can assess the work. What it will not do is state a price on a job it cannot see, because a figure on your site can become a binding expectation under UCC 2-313 when the real cost comes in higher. Fred books the visit and lets the crew quote what they can stand behind.
Why is the chatbot texting customers a legal risk?
Because automated calls and texts to customers are governed by the TCPA’s consent rules, and a bot that fires off outreach on its own can create unconsented contact the business is liable for. A compliant assistant follows the consent and follow-up rules you set rather than improvising marketing messages. Fred captures the lead and hands off; it does not run its own outbound campaign.
How is Fred different from a generic chatbot told not to quote prices?
A prompt instruction is something the model abandons when a homeowner rephrases "how much" in a way it did not anticipate. Fred enforces its boundaries at the system level, deciding what is allowed before it responds, so an invented quote or a committed timeline is never generated regardless of wording. That is the difference between a bot that usually deflects and one that cannot make the promise.