DEV Community

VoiceFleet
VoiceFleet

Posted on • Originally published at voicefleet.ai

Building an AI Receptionist for Tradespeople — What We Learned

The problem nobody talks about

Here's a stat that surprised me: tradespeople miss 38–45% of incoming calls. Not because they don't care — because they're literally up a ladder, under a floor, or driving between jobs.

For an electrician getting 10 calls/day, that's 4 missed calls. At €250 average job value with 15% conversion, that's ~€29,000/year in lost revenue. Enough to buy a new van.

What we built

We built an AI voice agent that acts as a 24/7 receptionist specifically for trades businesses. When a customer calls:

  1. AI answers within 2 rings with the business name
  2. Understands the problem ("my kitchen sockets stopped working")
  3. Captures: name, address, urgency, job details
  4. Checks the tradesperson's calendar for availability
  5. Books the callout and sends confirmation to both parties
  6. Texts/WhatsApps the tradesperson a summary

Total call time: ~90 seconds. Works at 2am on a Saturday when someone's RCD trips.

Technical decisions

Speech recognition was the hardest part. Irish accents + place names (Dún Laoghaire, Ráth Cairn) required significant tuning. We also had to handle:

  • Simultaneous calls — during storms, every electrician's phone rings non-stop. AI handles unlimited concurrent calls.
  • Urgency classification — "no power in the house" gets flagged differently from "I need a socket added"
  • Calendar integration — real-time availability checking, not just message-taking

Cost architecture: The AI costs ~€99/month. A part-time admin costs €1,200–€1,800/month and only covers 20–30 hours/week. The AI never calls in sick and handles 5 calls simultaneously.

What surprised us

  1. Emergency calls at weird hours are common — about 15% of all calls come between 9pm and 8am
  2. Callers prefer talking to the AI over leaving voicemail — 80% of people who hit voicemail just hang up
  3. The biggest competitor isn't other AI products — it's "just not answering the phone" as the status quo
  4. Tradespeople don't want an app — they want a text/WhatsApp summary they can glance at between jobs

Results so far

We're seeing Irish electricians capture 3–5 additional jobs per week they would have otherwise missed. For a solo operator, that's €750–€1,250/week in recovered revenue.

The use case generalizes well to plumbers, builders, locksmiths — anyone who can't answer the phone while working with their hands.


If you're building voice AI for niche verticals, I'd love to compare notes. The trades/SME space is massively underserved.

Full case study: voicefleet.ai/blog/ai-receptionist-electricians-ireland

Top comments (0)