If you run a small service business in the UK and you're tired of missing calls, you've got four real options for handling the phone. Voicemail (the default), a traditional answering service, an AI receptionist, or hiring an actual human. Each has trade-offs. This piece breaks down the maths and the practical reality of each one so you can pick what's right for your stage of business.
Option one: voicemail
Cost: zero, technically. But the real cost is the missed business. UK industry data suggests around 80% of callers hang up rather than leave a voicemail, especially for emergency-driven service categories like plumbing, electrical work and locksmiths.
Voicemail also has secondary problems: callbacks that take place hours later when the customer has already booked elsewhere, the time you spend triaging messages, and the reputational damage from "never returns calls" reviews. For most UK service businesses doing more than ten calls a day, voicemail is the most expensive thing on this list once you count properly.
When voicemail makes sense: you're a brand-new sole trader doing one or two jobs a week, calls are easy to handle in the evening, and you genuinely have spare capacity to chase callbacks.
Option two: traditional answering service
Cost in the UK: typically £1.50-£3 per minute, with monthly minimums starting around £80-£150. Companies like Moneypenny, AnswerConnect and Alldayphones operate in this space.
Pros: real human answering, can take messages, can route urgent calls. Some offer scripted booking handling.
Cons: the human answering doesn't know your business beyond what's in the script, so they handle questions like "do you do power flushing?" or "are you Gas Safe registered?" generically. Quality varies day to day depending on which agent picks up. Premium charges for nights and weekends. The cost can spike unpredictably during busy periods.
Real example: a busy plumber doing twenty calls a day with average call time of four minutes is using 80 minutes a day, or roughly 1,600 minutes a month. At £2 a minute, that's £3,200/month. £38,400 a year for someone who isn't quite as good as having a real receptionist.
When a traditional answering service makes sense: you're already paying enterprise prices because you're a larger firm with consistent overflow needs and you genuinely want a human voice on every call. Below that scale, the maths usually doesn't work.
Option three: AI receptionist (Poppy and similar)
Cost in the UK: typically £49-£349 per month for a dedicated, trained AI that answers in your business name, knows your services and prices, books appointments, and sends you summaries by text and email.
Pros: trained on your specific business so it can handle trade questions properly. Answers every call instantly, 24/7, no premium for nights or weekends. Pricing is predictable and dramatically lower than human alternatives. Modern AI voices are genuinely difficult to distinguish from human receptionists for short interactions.
Cons: doesn't handle complex multi-step conversations as gracefully as a great human. Edge cases (heavy accent, very poor line quality, customer in distress) sometimes need human handover. Setup takes 10-30 minutes (faster than hiring, slower than just plugging in voicemail).
When AI makes sense: you're a UK service business doing 5-50 calls a day, you can't justify a full-time hire, and the cost of missed calls outweighs the cost of an AI subscription. This is the sweet spot for most plumbers, electricians, salons, builders, locksmiths, gardeners and similar trades.
Real example: the same plumber from above (twenty calls/day) on Poppy's Pro plan at £179/month would cost £2,148/year. Versus £38,400 for the human answering service. Roughly 18x cheaper for similar quality on most calls. The savings alone fund a small marketing budget.
Hear how Poppy actually sounds
Real call, unedited. Decide for yourself whether you can tell she's AI.
Listen to a real Poppy call →Option four: hire a real receptionist
Cost in the UK: a full-time receptionist for a service business runs £22,000-£32,000 per year (London is higher). Plus employer NI, pension contributions, holiday pay, sick pay, equipment, software licences. True total cost is typically £30k-£42k a year.
Pros: a great human is still better than the best AI for genuinely complex conversations, customer relationship building, and handling edge cases. They know your customers personally over time.
Cons: cost. Recruitment time. Holiday and sickness cover. Training time. Out-of-hours gaps unless you also pay overtime. Many UK service businesses can't justify the spend until they're doing £400k-£500k annual turnover. Most don't have the management bandwidth to onboard properly.
When hiring makes sense: you're past £500k annual turnover, your call complexity is genuinely beyond what AI can handle (medical, high-value B2B, complex consultative selling), and you have the management time to recruit and train someone properly.
Side-by-side: a typical UK service business
Take a UK plumbing firm doing 100 calls a week, average ticket £180, currently with 25% of calls missed:
Annual cost (or saving) by option
For a business at this scale, AI receptionist wins on pure economics. As you grow past £500k turnover and your call complexity goes up, hiring becomes more justifiable. But for the first £500k of annual turnover, AI is the clear winner for almost every UK service business.
How to decide
Three quick questions:
- Are you missing more than five calls a week and wishing you weren't? If yes, do something. If no, voicemail is fine.
- Are you doing more than £500k a year and have complex consultative calls? If yes, hire a human. If no, an AI receptionist will do the same job at 5-10% of the cost.
- Do you specifically want a real human voice on every call? If yes, traditional answering service. Be ready for the cost. If no, AI is genuinely going to surprise you with how good it is in 2026.
For most UK service businesses below £500k turnover, the answer is AI receptionist. The technology has gotten good enough that the only reason to pay 20x more for a human is if you specifically want a human, not because the human will do the job better.
The only reason to pay 20x more for a human in 2026 is if you specifically want a human. Not because the human will do the job better.