Case Study: AI Voice Agent Cuts Call Costs from £24K to £4K
How we built an AI voice agent with ElevenLabs and Twilio that replaced a £2,000/month outsourced call center for a UK surveying firm — saving 83% on reception costs.
I just helped a surveying firm in the UK reduce their call reception costs from £24,000 per year to £4,000 using an AI voice agent.
The problem
The firm was paying Moneypenny £2,000/month for outsourced call handling. Professional, but expensive — and limited to business hours.
The solution: AI voice agent with ElevenLabs + Twilio
At HermesOps, we built a conversational AI voice agent that handles 100% of inbound calls:
- ElevenLabs for natural-sounding text-to-speech — the voice is indistinguishable from a real receptionist
- Twilio for telephony infrastructure — call routing, recording, and SMS notifications
- OpenAI GPT-4 for conversation logic — understanding caller intent, asking the right follow-up questions
- n8n for workflow automation — routing qualified leads to the right team member via email
- HubSpot CRM integration — every call is logged with contact details, service requested, and urgency
What the agent does
- Answers every call like a live person — 24/7, no hold times
- Determines what service the client needs (boundary survey, topographic, measured building)
- Collects project details, location, and deadlines
- Validates contact details
- Sends a structured ticket to the team via email
The email looks like this:
“John Smith needs a boundary survey for a property sale in Reading. Deadline: 3 weeks. Phone: 077XX”
A qualified lead ready to work.
The numbers
| Metric | Before (Moneypenny) | After (AI Agent) |
|---|---|---|
| Monthly cost | £2,000 | £60 |
| Annual cost | £24,000 | £4,000 |
| Availability | Business hours | 24/7/365 |
| Cost per call | ~£10 | £0.30 |
| Lead capture | During hours only | Evenings, weekends, holidays |
The tech stack
- Voice AI: ElevenLabs (TTS), Deepgram (STT)
- Telephony: Twilio
- LLM: OpenAI GPT-4 for conversation
- Automation: n8n for workflows
- CRM: HubSpot integration
- Hosting: Docker on a VPS
Where AI voice agents apply
This same architecture works for any business that handles inbound calls:
- Law firms — screening consultations, conflict checks, intake forms
- Real estate agencies — qualifying buyers, booking viewings
- Recruiting firms — filtering candidates, scheduling interviews
- Medical practices — patient triage, appointment booking
- Trades & services — plumbers, electricians, HVAC — capturing leads after hours
ROI is instantaneous. The agent paid for itself in the first week.
After phased testing, this agent went live handling 100% of inbound calls — capturing two qualified leads on the first day.