CASE STUDY · FIELD SERVICES AI (SMB)
Automating inbound call handling for a home services business.
- 82%82% of routine calls handled without staff
01 · PROJECT OVERVIEW
The client and the brief.
An anonymised home services business. Name withheld under confidentiality.
A voice agent with scripted flows and hand-off to staff. The scripts cover the common reasons people call: booking a visit, asking about a service, rescheduling and reporting a problem. Anything outside those flows goes to a person.
- 1After-hours calls were lost.
- 2The business had a small office team who left at the end of the day, and an on-call technician who could not always pick up.
- 3Customers with a non-urgent problem called elsewhere.
- 4Customers with a real emergency sometimes waited too long to reach someone.
02 · WHAT WENT WRONG
What went wrong, and how we fixed it.
Every engagement hits problems. These were the ones that mattered here, and what we changed so they did not come back.
- PROBLEM 01Some callers hung up when they realised they were speaking to an automated voice, particularly older customers
We adjusted the opening line so it was short and clear, added an option to say "speak to a person", and shortened the first questions. Drop-offs in the first thirty seconds fell. We also learned to keep the voice pace slower, which improved comprehension on poor mobile lines.
- PROBLEM 02We also learned to keep the voice pace slower, which improved comprehension on poor mobile lines
03 · THE PROCESS
How we ran it.
- 01Week 1
We reviewed call recordings and notes with the owner. We listed the five or six reasons for most calls and the signals for an emergency.
- 02Weeks 2 to 3
We wrote and tested the flows, including awkward callers and unclear speech.
- 03Supervised launch
The agent took after-hours calls only. Staff checked a sample of transcripts each morning.
- 04Expansion
Once the team trusted it, we extended the hours it covered and tuned wording based on real calls.
04 · RESULTS
What changed.
Measured against the way the work ran before the engagement began.
- NOW 01
The owner checks a short morning list of calls and bookings instead of listening to voicemails.
- NOW 02
The on-call technician gets a written summary with every transfer.
- NOW 03
New call types are added when the transcripts show a pattern, not by guesswork.
- NOW 04
The team treats the agent as part of the front desk and updates its rules as services and prices change.
- 82%
82% of routine calls handled without staff
Urgent calls transferred to an on-call technician
05 · WHAT WE BUILT
How it works.
- Scripted flows with defined steps, rather than open conversation.This makes behaviour predictable and easy to review.
- A language model handles the wording and unexpected phrasing, but cannot add steps, quote prices or promise outcomes outside the script.
- Calls that match urgent criteria are transferred to the on-call technician.If the technician does not answer, the agent tries again and sends a message with the details.
- Routine bookings are saved as requests or booked into free slots, depending on the owner's preference.
- Transcripts and summaries are stored and linked to the customer record.
- The agent always offers a route to a person, and says what will happen next.
- Controls and handover
- The agent cannot take payment, quote a fixed price or confirm an arrival window that the schedule does not support. It reads back the address and phone number before ending the call, because a wrong digit is the most common cause of a lost job. We reviewed a sample of calls with the owner each week for the first month, then moved to a monthly review.
06 · QUICK ANSWERS
The short version.
- What problem did the client have?
- After-hours calls were lost. The business had a small office team who left at the end of the day, and an on-call technician who could not always pick up. Customers with a non-urgent problem called elsewhere. Customers with a real emergency sometimes waited too long to reach someone.
- What did Brandsmashers build?
- A voice agent with scripted flows and hand-off to staff. The scripts cover the common reasons people call: booking a visit, asking about a service, rescheduling and reporting a problem. Anything outside those flows goes to a person.
- What were the results?
- 82% of routine calls handled without staff; urgent calls transferred to an on-call technician.
YOUR TURN
Have a similar requirement?
Tell us what is stuck. We will come back with a plan and matched, pre-vetted engineers.
