Contractors · AI receptionist
An AI receptionist for contractors that takes the job details, not the booking
A homeowner calling about a remodel wants to describe the work and ask a few questions. We've built and tested a GoHighLevel (GHL) voice agent and chat agent that do exactly that, save what they hear on the contact record, and stop there. No price, no booking, and nothing goes live until you have tested it.
The job
Intake and answers, and nothing past that
The AI's part ends where your estimator's begins.
Most of what a homeowner says on a first call is the brief for your site visit: what they want done, which part of the house, and roughly when. Then come the questions: do you take this kind of work, and how does your process run? An AI can handle both halves of that conversation. What it should not do is put a number on work nobody from your company has seen.
For a general contractor we built a voice agent and a chat agent in HighLevel with that line drawn in from the start. The owner set the scope: collect information and answer questions, and nothing more. Neither agent gives a price, and neither books an appointment, a call or a visit. That was the owner's decision for the business, made on purpose and open to change later.
What the agents hear lands on the contact. Both agents write what they collect into custom fields on the contact record, so whoever calls the homeowner back reads the job in one place, whether it came in by phone or by chat. Those fields are not private to the AI. The 90-day lead reactivation runs on them too, which is why every custom field went across when the build moved into the contractor's main account.
One approved source for every answer. We built a knowledge base from the contractor's website and put it through approval. After that, the approved knowledge base became the source the prompts and answers are written from. We do not hand the raw website to every prompt by default; the rule is that copy comes from the approved version, so the sourcing stays the same everywhere. A caller and a person typing in chat get their answers from the same approved wording.
Built in one account, switched on in another. The agents were built in a separate build account, and the owner tested them there by calling and chatting. A snapshot then carried them into the main account, where they sit switched off until the owner tests them again in the account they will actually run in. No real homeowner reaches either agent before that second test.
Why stop short of booking? In remodeling the step after a first call is usually someone going out to see the job, and that visit is best set by a person who knows what the job needs and who should go.
What we set up
The parts of the intake build
Built from the plan you signed, first in a build account, then in yours.
- A HighLevel voice agent that takes down the job and answers questions
- A chat agent with the same scope, for people who would rather type
- Custom fields on the contact for everything the agents collect
- A knowledge base built from your website and approved before it is used
- Prompts written from that approved knowledge base, not from the live site
- No pricing in what the AI is given to say, by the owner's choice
- No booking step in either agent
- A separate build account for testing, then a snapshot into your own account, switched off
How a setup runs
From a first call to an agent you've switched on
A call with Rob
Thirty minutes on what homeowners tell you on a first call, what they tend to ask, and what your team needs written down before anyone goes out to see a job.
Every line in writing
The greeting, each question the AI asks, every knowledge-base answer and how it closes a conversation are in the plan document word for word. You sign it before we build.
Build and test in a separate account
You call the voice agent and chat with the chat agent in the build account, and we adjust until both stay inside the scope you set.
Move it into your account, switched off
A snapshot carries the agents and their custom fields into your HighLevel account. They stay off until you have tested them there and told us to go.
QA report, then changes
At handover you get a QA report with screenshots of each part. A later change starts with a verdict: covered by the plan you signed, or new scope.
What you bring
What we need from your team
Your first-call questions
What your office asks a new caller, and what your estimator needs to know before a site visit. Those become the questions the AI asks.
Your website
We build the knowledge base from it. You approve the result before a single prompt is written from it.
Where your notes live
If your team keeps scope, measurements and timelines in contact notes, tell us. We check what the AI can and cannot read in your account before the plan is signed.
When the AI should answer
Which hours or days it covers, and how long your own people get first. You decide that before anything is switched on.
Where it stops
The AI does not quote, estimate or hint at a price, and it books no site visit, no call and nothing else. It collects the job details, answers from the approved knowledge base, and leaves every next step to your team.
Services used
Where the two agents get set up
HighLevel Voice AI agents
Inbound and outbound voice agents on HighLevel's own builder, with transfers, booking and the consent rules outbound calling needs.
See Voice AI agentsConversation AI and DM bots
Chat bots for texts, web chat and social DMs that answer from your knowledge base, book the appointment and then go quiet.
See Conversation AI setupQuestions
What contractors ask about an AI receptionist
What if a caller pushes for a rough price?
Can the AI put the site visit on our calendar?
Do the phone agent and the chat agent say the same things?
Will it go live the day it's built?
Does it replace the person who answers our phones?
Who you talk to
Rob Eason
Rob runs HLAI Experts and takes the 1:1 calls himself.
He is a HighLevel Certified Admin with the second-most reviews in HighLevel's Certified Admin Directory worldwide, all of them 5 out of 5.
He won HighLevel's Top New Subaccounts via AI Employee award in its 2025 Summer of AI Awards and taught HighLevel's official masterclass on its AI Employee.
Before HighLevel, he co-owned a 60-employee financial services firm and sold his share.
On the call he looks at your account, not a script.
5/5 across 83 reviews in HighLevel's Certified Admin DirectoryHear what your AI receptionist would ask
In 30 minutes Rob walks through a first call at your company and shows which details the AI would take down, what it would answer, and where your team takes over.