AI Lead Qualifier
An AI voice receptionist that answers the calls a busy tradesperson would miss, qualifies the job, and books it, before the caller hangs up.
A working AI voice agent for trades businesses, built solo. When the owner cannot pick up, it answers in seconds, works out what is wrong and how urgent it is, and books the job straight into the calendar. A proof of concept for exactly the kind of AI a small business can actually use.
Small trades operators lose real money to missed calls. They are on a roof or under a sink during the working day and physically cannot answer the phone, and a missed call is usually a missed job that goes to the next contractor on Google.
Voicemail does not fix it, because urgent callers do not leave messages. They hang up and dial the next number. I wanted to see how close modern AI could get to a great human receptionist for a one-person business.
AI Lead Qualifier is a voice receptionist built around a demo plumbing company. When a call comes in, an AI agent picks up in about two seconds and holds a natural conversation, then does three things a generic chatbot cannot:
- Qualifies the lead: it captures the caller's name, the issue, the location and the property type, not just a message
- Triages urgency: a burst pipe or a gas leak is flagged and sorted to the top of the inbox, ahead of a price question
- Closes the loop: it checks real calendar availability, books the appointment live, then emails the owner a transcript
It is a complete, deployed proof of concept, built around a mock trades business with its own knowledge base, pricing and service area, plus a website chatbot that does the same job in text. I built it to prove the concept end to end, then parked it. There are no real users behind it, the point was to show the pattern works.
It is genuinely agentic, not a script. The AI has live tools it can call mid-conversation: check the calendar, book the appointment, and search the business's knowledge base, all inside a roughly two-second response budget. The instructions are assembled per business, injecting its services, pricing, hours and escalation rules, so a gas leak is fast-tracked with minimal questions.
Under the hood it stitches together a modern AI voice stack: a voice-orchestration layer, Claude as the reasoning brain, ElevenLabs for the voice, Deepgram for transcription, and calendar, SMS and email integrations to close the loop. The point was not any single tool, it was wiring them into one agent that reliably does a real job.
This is the shape of AI most small businesses actually need: not a chatbot bolted onto a website, but an agent that does a complete job, answer, qualify, act, and hands the owner only what needs a human. I can scope and build exactly this kind of practical AI for your business.
Want something like this built for your business?
I turn real problems into working products, fast, with AI. Let’s talk.
Start a conversationStart a conversation