Missed calls cost UK businesses more than they realise. Not just lost enquiries but delayed bookings, frustrated patients, and opportunities that quietly go elsewhere.
An AI receptionist is designed to stop that happening. It answers calls instantly, handles routine requests, books appointments, routes callers correctly, and works around the clock — without the overhead of a full-time reception team.
This page explains what an AI receptionist is, how it works in the UK, when it makes sense, and how to choose the right option, with clear links to deeper comparisons and pricing where appropriate.
For UK SMEs, AI receptionist software typically costs £150–£400 per month depending on call volume, logic depth, and integration requirements. At current UK wage levels, most organisations reach break‑even after 5–15 captured enquiries per month, making this one of the fastest-payback operational technologies available.
Here is the reality in pricing terms. The market clusters into three tiers:
Entry Tier (£100–£180/month): Basic call answering, message taking, simple routing. Best for low‑risk, low‑volume environments.
Mid Tier (£180–£300/month): Action‑taking AI that books appointments, writes to CRMs, and handles moderate complexity. This is where most UK clinics and professional services land.
Enterprise Tier (£300–£600+/month): Context‑aware agents with safeguarding logic, escalation paths, and compliance controls. Necessary for healthcare, legal, and high‑value inbound sales.
Against a fully loaded UK receptionist cost of £32k–£40k per year, even premium AI solutions typically undercut human coverage by 70–85%, before considering out‑of‑hours capture and zero sick leave.
Want to Know How Much You Could Be Losing By Missing Out Of Office Calls?
The conversation usually starts in the wrong place. It starts with features. The correct place to start is the cost of a missed call.
In UK service businesses, missed inbound calls convert poorly when followed up. Industry data across healthcare, legal, and home services consistently shows that 30–45% of missed callers never re‑engage. For a dental practice, that might be a £60 hygiene appointment. For a law firm, it might be a £2,500 matter. The variance is wide, but the leakage is constant.
Now anchor this against the human baseline. A full‑time receptionist in the UK earns £27k–£35k. Add employer National Insurance at 13.8%, pension contributions, holiday cover, sickness, and management overhead, and the fully loaded cost sits closer to £32k–£40k per annum. That is before you deal with lunch breaks, evenings, or Mondays at 8:59am.
Here is where the math breaks. If an AI receptionist is £250 per month thats £3,000 per year. Even a more advanced £400 per month system lands at £4,800 per year. The decision isn’t marginal, but even an out of hours AI Receptionist could pay for itself in month 1.
The break‑even calculation is therefore simple. Assume a conservative £150 gross margin per captured enquiry. At £250 per month, the system needs to convert two additional enquiries per month to cover its cost. Everything beyond that is pure operational upside.
This is why Operations Directors care less about whether the AI sounds “human” and more about whether it answers on the third ring, never forgets to log a call, and does not call in sick during half‑term.
UK pricing is converging, but the billing models still matter because they change risk.
The flat‑rate model, used by providers such as Dialzara and NextPhone, typically sits between £150 and £250 per month for what is marketed as “unlimited” or “high‑volume” usage. In practice, fair‑use thresholds apply, but for most SMEs this model offers predictability. You know your monthly exposure, and Finance can forecast with confidence.
The per‑minute model, common with Synthflow deployments or Vapi‑based custom builds, ranges from £0.10 to £0.40 per minute depending on voice quality, latency, and language models used. At low volume this looks cheap. At scale, it can quietly exceed flat‑rate pricing. A practice taking 2,000 minutes of calls per month at £0.25 per minute is already at £500, before platform fees.
The hybrid or legacy model, seen with providers like Smith.ai and Moneypenny, combines a base subscription with per‑call or per‑minute charges. UK pricing often lands between £200 and £500 per month once real usage is accounted for. These services blur the line between AI and human support, which is reassuring for some buyers but expensive if 90% of calls are routine.
Hidden costs are where UK buyers get caught. Number porting for 01, 02, or 03 numbers can carry one‑off fees of £20–£50, and some providers quietly add monthly hosting fees for UK numbers. VAT matters. If you are not VAT‑registered, that extra 20% is real cash, not reclaimable. Overage penalties are another pressure point. Some flat‑rate plans throttle performance or charge punitive rates once thresholds are exceeded, which tends to surface only after a marketing campaign spikes call volume.
The monthly fee is only half the story. The number that matters is what’s included before overages, bolt-ons, or “fair use” clauses start doing damage. Here’s how several popular providers structure allowances, so you can compare like-for-like.
| Provider | Entry Tier Allowance | Mid-Tier Allowance | High-Tier Allowance | Overage / Notes |
|---|---|---|---|---|
| Dialzara | 60 minutes | 220 minutes | 500 minutes | Overage listed as ~$0.48/min (USD pricing) |
| CyberStaff | 50 calls | 150 calls | 500 calls | Per-call overage (varies by plan); optional human handoff add-on |
| My AI Front Desk | 200 minutes/month | 300 minutes/month | Custom / volume discounts | Extra usage shown as “extra credits”; minutes are the core metered unit. |
| CloudTalk | “Unlimited calls” to US/Canada + inbound/internal included* | Country bundles shown (e.g., outbound minutes per user/month in certain countries) | Varies by plan + add-ons | Allowances depend on destination/country bundle; check per-country included minutes. |
| Moneypenny | Varies (quote) | Varies (quote) | Varies (quote) | Typically tailored around coverage + call handling needs |
| Norango | Varies (quote) | Varies (quote) | Varies (quote) | Usually bespoke (human + AI mix), so allowances are defined in the proposal |
*Even when a provider markets “unlimited”, treat it as a billing model, not a blank cheque. Always check the fair usage language and which destinations/counties are included.
Quick reading guide (so you don’t get misled by the labels):
The most common mistake UK SMEs make is choosing a plan based on the monthly subscription fee rather than their actual call data. If you receive 200 calls a month but sign up for a 50-call “Starter” plan because it looks cheap, your overage charges will likely exceed the cost of the “Pro” tier within the first fortnight.
Before committing, audit your last three months of phone bills. Look for:
Not all “AI receptionists” are doing the same job. Price is a proxy for logic depth.
At the £ level, typically under £100 per month, you are buying a scripted responder. It can answer FAQs, take a message, and route calls based on keypad input. This is voicemail with better manners. It reduces interruptions but does not materially change outcomes.
At the ££ level, usually £180–£300 per month, the agent becomes an action‑taker. It can book into Cliniko or Jane, create or update records, take deposits, and integrate with systems like ResDiary. This is where operational leverage appears. Calls stop being interruptions and start being transactions.
At the £££ level, north of £300 per month, context enters the equation. The agent understands urgency, recognises safeguarding keywords, handles escalation paths, and performs warm transfers when required. For healthcare and legal environments, this is not optional. It is risk management disguised as software.
This is why £30 per month bots exist and why they disappoint. They answer questions. They do not run a front desk.
Most failed deployments do not fail technically. They fail economically.
One common pattern is paying for “human‑in‑the‑loop” oversight when AI could resolve the call entirely. If 90% of calls are booking, rescheduling, or basic enquiries, paying per interaction for a human fall-back can erode the maximum cost advantage but triage will always be needed and it may be prudent to have human handover in an “emergency”.
Latency is another silent killer. Systems with poor speech turn‑taking cause talk‑over, stretching calls to 30 seconds that still bill as a full minute. On per‑minute pricing, this is death by a thousand cuts.
Then there is the integration tax. Some providers charge extra for CRM write‑access, appointment creation, or even basic webhook usage. What looks like a £200 plan quietly becomes £350 once it actually does the job you assumed it would.
The calm audit question is always the same: “What am I paying for that does not directly increase captured demand or reduce staff load?” If the answer is unclear, buyer’s remorse follows.
Yes. For VAT‑registered businesses, AI receptionist software is treated as a standard SaaS operating expense, and VAT can be reclaimed in the usual way. For non‑VAT‑registered clinics and practices, VAT is a real cost and should be factored into headline pricing comparisons.
Most modern providers charge £0–£500 for setup. Flat‑rate platforms like Dialzara and NextPhone often include setup, while bespoke Vapi‑based builds may charge more for call‑flow design, compliance logic, and testing.
Flat‑rate providers usually apply fair‑use thresholds and may throttle or upsell if volumes remain elevated. Per‑minute models simply bill the usage. The financial risk sits with you, not the vendor, which is why Finance teams prefer predictability.
In most cases, yes. UK geographic numbers can be ported, but expect one‑off porting fees and a short transition window. Always confirm whether the provider charges ongoing hosting fees for your number, as this varies.
The reality is that AI receptionist pricing is no longer the strategic question. Unit economics are. At UK wage levels, the payback period is measured in days, not quarters, provided the system actually answers calls, takes action, and integrates cleanly.
Operations Directors do not buy AI because it is clever. They buy it because it closes leakage in the most expensive part of the funnel. When priced and deployed correctly, the front desk stops being a cost centre and starts behaving like infrastructure.
That is the frame that survives procurement, audit, and the next budget cycle.