Austin Phone Service answers the call, captures the lead, flags urgency, and alerts the owner in real time. Built for handyman, HVAC, plumbing, electrical, and other call-driven trades.
Prompt Engines · Austin, TX · March 2026
When the phone rings, the owner is often driving, on a ladder, under a sink, or with another customer. The first business to respond usually wins. Everyone else bleeds demand.
If the system recovers even one missed HVAC, plumbing, or electrical job per month, it often pays for itself. The product is not “AI.” The product is fewer lost calls and faster owner follow-up.
The wedge is local, call-driven, relationship-heavy service businesses. Austin gives the product a tight geography, a recognisable local number, and fast feedback loops.
A 512/737 number and Austin-specific language make the experience feel local, not outsourced. That matters when the customer needs someone at their house.
Austin has a large pool of owner-operated trades businesses that live on inbound calls but are too small for full reception staff.
Prompt tuning, call reviews, and owner interviews can happen in person. That shortens the loop between product and market.
Start with a local services wedge. Expand only after the pilot proves the intake, alerting, and follow-up loop.
The receptionist's job is narrow by design: capture the lead, classify urgency, and set up the owner callback. That keeps the system safe and useful.
If the caller reports a gas leak, flooding, electrical fire, or another safety emergency, the agent tells them to call 911 first and separately flags the lead for the owner.
If the caller is abusive or the call is clearly spam, the agent ends politely. This protects owner time and reduces junk records.
The system is not trying to close the job on the call. It is trying to produce a clean, actionable lead record the owner can trust and work from immediately.
The current prototype already connects telephony, real-time conversation, API ingestion, storage, and owner notifications.
If the caller disconnects before the full lead is captured, the scenario can still send a fallback notification so the owner knows the call happened.
The API validates required fields before storing the lead. Required fields include caller name, service need, urgency, business ID, and caller ID number.
Austin Phone Service comes with an admin surface designed for the person actually running the business: fast status changes, urgency visibility, recordings, and clear follow-up.
Each record stores caller name, phone, caller ID number, service needed, address, urgency, notes, call duration, recording URL, session ID, and status.
Trades owners do not want another abstract AI dashboard. They want to see who called, how urgent it is, and whether someone followed up. This surface keeps the workflow grounded in operations.
The economics work because trades jobs are valuable and the system cost is low relative to even a single saved lead.
Austin pilot: first 3–5 local businesses can run at reduced pilot rates in exchange for call reviews, case studies, and prompt-tuning feedback.
The broader voice-agents model in this repo targets roughly 82–86% gross margin, with API spend low relative to subscription revenue.
You do not need dozens of saved calls. You need a small number of missed leads recovered every month. That is enough to justify the spend.
The repo is not just a deck. It already includes the call scenario, the API, the schema, the notification path, the dashboard, and tests.
The first goal is not scale. It is to prove that the system answers real calls, captures usable leads, and helps the owner respond faster than before.
Do not start with dozens of accounts. Start with a few owners who actually care, review the recordings, and improve the script against live traffic.
Owners trust the alerts, callers finish the conversation, and the business wins jobs it would have otherwise missed.
Austin Phone Service is a local wedge for a broader voice-ops business: a receptionist that answers the phone, captures structured leads, and gives the owner a usable pipeline instead of a voicemail graveyard.
Prompt Engines · Austin Phone Service · 512-first pilot