How a plumbing company never missed another emergency call
Emergency calls were going to voicemail any time the office was on a job. After Boafo Agent took over inbound on phone, web and WhatsApp, the company captured 100% of emergency calls, booked 54% more jobs, and retired a GBP 1,400-a-month answering service.
Results at a glance
The problem
The company runs five engineers and an office manager who also handles inbound calls. Whenever the office manager was on the phone with a customer, the next caller went to voicemail and often never called back.
Emergencies were the worst case. Boiler failures, burst pipes and blocked drains do not wait for the office to be free. The owner estimated the business was losing two to three emergency call-outs every weekend, each worth an average of GBP 280 plus the lifetime value of a new customer who would have stayed for annual servicing.
An outsourced answering service had been in place for two years at GBP 1,400 a month. The agents took messages but could not check the diary, could not confirm whether an engineer was nearby, and routinely promised callbacks that the office could not actually honour, which generated 1-star reviews.
The owner wanted three things: an instant answer on every inbound call, a real diary check before any promise was made, and a clean handoff to the on-call engineer when the job was genuinely an emergency.
The solution
Boafo Agent was deployed on the company main inbound line, the website chat widget and a WhatsApp business number. The AI was trained on the service area, current pricing for standard jobs, the engineer rota and the rules for what counts as an emergency.
The emergency flow is the heart of the deployment. The AI asks two qualifying questions, reads the postcode, checks which engineer is on call and within the service area, then texts that engineer with the address, fault description and the customer callback number. The customer is given a confirmed ETA window before they hang up.
Standard jobs follow a separate flow. The AI quotes the fixed call-out fee, offers the next two available diary slots that match the engineer geographic route, and books the job directly into the engineering calendar with a confirmation text.
Out-of-hours behaviour is explicit. Between 20:00 and 07:00 only genuine emergencies escalate to the on-call engineer. Standard requests are captured in full and booked into the next available morning slot, so the office walks in to a triaged diary instead of a voicemail backlog.
The results
Since launch the company has not missed a single emergency call. Every inbound is answered in an average of 6 seconds, day or night.
Total jobs booked rose 54% in the first 90 days. The bulk of the lift came from previously-lost out-of-hours and overflow calls, not from new marketing spend.
The outsourced answering service was retired in month two, freeing GBP 1,400 a month and removing a recurring source of 1-star reviews about broken callback promises.
5-star reviews rose 71% in the first quarter, driven by faster pickup, accurate ETAs and the fact that customers now spoke to something that knew the diary in real time.
“We used to lose a boiler emergency every other Saturday because no one was near the phone. Now the AI picks up in seconds, takes the address, checks the diary and texts the on-call engineer. We have not missed a single emergency since the day it went live.”
Before vs. after
| Metric | Before | After Boafo Agent |
|---|---|---|
| Missed emergency calls / month | 8 to 12 | 0 |
| Average pickup time | 4 rings, often to voicemail | 6 seconds |
| Out-of-hours capture | Voicemail only | 100% live AI capture |
| Diary accuracy on bookings | around 70% (manual) | 100% (real-time check) |
| Answering service cost | GBP 1,400 / month | GBP 0 (retired) |
| Jobs booked / month | Baseline | +54% |
| 5-star reviews / quarter | Baseline | +71% |
Implementation playbook
The owner ran the deployment in three deliberate phases. Phase one was the website chat widget and WhatsApp only, on a soft launch, with the existing answering service still in place. The goal was to test the booking and emergency flows in low-stakes conditions and build the engineers confidence that the AI would not embarrass them with a misrouted call.
Phase two switched the main inbound phone line to the AI during business hours, with the office manager listening live to the first three days of calls and intervening only if the AI got something materially wrong. Across those three days the office manager intervened twice, both times to correct a postcode the customer had mumbled.
Phase three retired the answering service and moved the out-of-hours line to the AI. The owner deliberately ran both in parallel for three weeks before flipping the switch, so the comparison numbers were unarguable. The AI captured 100% of calls versus 84% for the outsourced agents, and booked 31% more emergency jobs because it actually knew which engineer was on call and where.
The engineers were the surprise advocates. They had been the most sceptical group before launch and became the strongest internal champions within a fortnight, because the emergency texts they now received arrived with full address, fault description and customer callback number, instead of paraphrased voicemails from offshore agents. Average time-to-arrival on emergencies dropped 23 minutes.
The owner now uses the Boafo Agent dashboard as a weekly operations review. The metrics that matter most are missed-call count (target zero), emergency response time (target under 30 minutes) and 5-star review rate. Any week that any of those three drifts triggers a small flow review the following Monday.
What is next
The owner now treats the AI as the front door of the business and the office manager as the second line. New engineers are added to the rota in the AI dashboard the same day they start, and pricing updates take seconds.
Next on the list: a planned-maintenance reminder flow that proactively books annual boiler services for existing customers, currently the single largest source of unrealised revenue.