AI Receptionist for UK Tradesmen: The Honest 2026 Buyer's Guide

If you run a UK trade business, the phone is how work comes in. It's also one of the things you can't reliably answer — because you're under a sink, in a loft, up a ladder, or driving between jobs.

Here's the honest cost of that.

A 2025 Paperclip survey of 142 UK SMEs found 47% of initial calls go unanswered. UK call-tracking data published by Aira puts the figure at 62% across all UK SMEs. Of the people who hit voicemail, Invoca's home-services research shows 85% never try a second time. Across the whole UK economy, Paperclip estimates the loss at £30bn a year.

For one solo plumber averaging 30 calls a month at a £180 average callout, the number that matters is roughly £334 a month walking out via voicemail — and that's before the long-tail of "they hired the next plumber on Google so I never see them again."

This guide is for the people who've finally had enough of it. We'll lay out every UK AI receptionist option — the real companies, the real prices, the real trade-offs — and show you exactly how to evaluate them. Disclaimer up front: we're Invox AI. We sell one of the products in this market. We mention it once at the end, and we'd rather you pick the right tool for your business — even if it isn't ours — than buy ours and hate it.

What an AI receptionist actually is (and how it differs from voicemail and a human PA)

The category is muddy because three different things all call themselves "answering services":

The thing tradesmen often miss: most "AI receptionists" don't actually book jobs. They answer the phone, capture the details, and text you a message. That's a fancy voicemail. The product you want is one that sees a real availability slot in your calendar, offers it to the customer, gets a yes, and creates the booking — without you tapping anything afterwards. That's the line that separates the ones worth paying for from the ones that aren't.

The maths: how much UK trades actually lose to missed calls

The maths is dull but worth doing properly because it tells you exactly what the upside is worth.

Inputs (typical UK solo plumber):

• 30 inbound calls per month
• 62% miss rate (Aira)
• 85% of voicemail callers never try a second time (Invoca)
• 10% close rate on answered enquiries (industry-standard for cold callers)
• £180 average callout

The arithmetic: 30 × 62% = 18.6 missed calls. 18.6 × 85% = 15.8 leads gone for good. 15.8 × 10% close = 1.58 lost jobs × £180 = £284 a month in this conservative version. Layer in the higher-value calls (emergency, install jobs, repeat customers) and most plumbers we've modelled land between £300-£500 a month.

The other side of the same coin is first-responder advantage. A Harvard Business Review study analysing 2.24 million sales leads found firms that responded within an hour were nearly 7× more likely to qualify the lead than those who waited even 60 minutes longer. Other research finds 78% of customers buy from whoever responds first, and responding within 5 minutes makes you 21× more likely to qualify the lead.

The implication for trades is bleak: when you don't pick up, the customer's already on the next plumber's website. Missing the call doesn't just lose that job — it gives the job to your direct competitor. AI receptionists are the only economical way to be that first responder 24/7 without paying a human to sit by a phone overnight.

The four jobs an AI receptionist must do (and why most fail at #4)

1. Pick up

Sounds trivial. Isn't. The AI receptionist needs sub-two-second pickup latency, must handle UK numbers reliably (some US-built systems struggle with +44 forwarding), and shouldn't drop calls during peak hours. Test this before you commit by ringing the system at 03:00 on a Saturday. If it takes 4 rings or routes to voicemail anyway, walk away.

2. Sound human

Modern AI voice tech is genuinely close to indistinguishable from a polite human. The good systems use voice models from providers like ElevenLabs or Cartesia and tune them for British accents. The bad ones still have telltale robotic prosody. Listen carefully on a call before signing up — most providers offer a demo number.

3. Qualify the job

This is where trade-specific tuning matters. A generic AI receptionist trained on dental practices won't know to ask "combi or system boiler" before quoting a callout. The trades-targeted options — ARROW, CallChimps, TradeMore, Invox — claim explicit trade training. The multi-sector tools (Voqal, Team-Connect, the human PAs) don't, which means more calls end with "I'll have someone call you back" instead of a confirmed booking.

4. Actually book it

The line in the sand. The receptionist's job is not to take a message — it's to put a job in your calendar. Before you commit to any provider, ring it pretending to be a customer, ask for a slot at "10am tomorrow," accept the offer, and then check whether the booking has actually appeared in your Google Calendar (or whichever calendar you use). If it hasn't, you're paying for an answering machine with a posh voice.

The UK landscape (who's actually shipping)

A few things worth flagging:

• ARROW is genuinely good at one thing. Founder Ben Milsom comes from a marketing background (LinkedIn) and runs a tight content operation. Their writing is the best in the UK market. Their product covers call answering well; it doesn't ship a CRM, calendar two-way booking, or SMS reply — those are separate tools. £99/mo for 150 minutes is the right price if you only need answering.
• TradeMore is the dark horse. Vetted Ltd has been around since 2002, and the broader product (AI receptionist + quotes + invoices + calendar + CRM) overlaps closely with Invox. Worth watching but pricing isn't public.
• Voqal AI is too new to bet on. Incorporated March 2026 means you'd be a foundational customer of an unproven 7-week-old company.
• Moneypenny's AI tier is real but priced for office-based businesses. If you're a solo trade, the £100+/mo entry doesn't compete with the trades-targeted tools.
• The US imports (Marlie at $49-199, Trillet at $49, Allo at $45) all rank in UK SERPs. They don't ship UK numbers natively and don't tune for UK trade workflows — the cost saving rarely covers the operational pain.

Real outcomes: what UK trades actually report

We say "directional" because every number in this section is self-reported by customers to vendors, which is not the same as audited evidence. Read them as plausible upper-bound outcomes, not guaranteed results.

ARROW publishes case studies (aiphonecalls.co.uk/case-studies) with named UK customers including:

"I've not seen anything as good as what you've put together. Compared to something like a traditional phone answering service, it's just so cost-effective." — Matt Billingham, Billingham Properties (Stourbridge). ARROW reports 1,000+ calls handled in 9 months with zero missed and £1,500+/month savings vs traditional answering services.

"Over 50 calls came in during my week-long holiday. Each was answered instantly and professionally. Not a single lead was lost." — Josh Gill, Carpet Cleaning Co (Ashford, Kent).

"Rose doesn't just answer the phone. She gives me back my time, my focus — and my evenings." — Jack Hale, Heartwood Tree Company (Tendring).

Other named ARROW customers (Be Sure Repairs, The Sussex Sign Company, Philip Raby Specialist Cars, ASAP Repairs, Hunt Visuals) report similar qualitative wins.

Industry-wide claims: secondary research from US-based plumbing-vertical writers (Trillet, Marlie) puts the conversion lift at 25-40% more captured leads vs voicemail and 10+ hours per week of admin time saved. UK-specific audited numbers don't exist yet — the industry is too young — but the directional case is consistent across vendors.

What's not yet proven publicly: long-term retention impact (does the customer who books with an AI come back as a repeat?) and call-quality impact on review scores (does an AI-handled job get the same star rating as a human-handled one?). If you want certainty on those, you'll have to wait another 12-18 months for proper third-party studies. Most trades who can't wait that long are running their own A/B test on a 14-day free trial.

How to evaluate any AI receptionist before you sign up (the 5-test method)

Run these on every option you're seriously considering. They take 10 minutes each, cost nothing, and will surface the differences faster than any sales call.

Test 1 — The 02:30 latency test

Ring the demo number at 02:30 on a weekend. Time the pickup. Anything over 3 rings, or anything that drops to voicemail, is automatically out. Most UK options should pick up in under 2 seconds. The ones that don't usually have routing or capacity issues you'd discover in production.

Test 2 — The trade-knowledge test

Ask a question only a real tradesman would ask. "How much for a Vaillant Ecotec Plus diagnostic in Salford on a Sunday?" or "Do you do Part-P notifiable jobs?" or "Will the engineer be Gas Safe registered?". The AI shouldn't make up a price, but it should show it understands the question. If the response is "I'll need to take a message", the system isn't trade-trained — you're paying for generic call handling.

Test 3 — The booking test

Ask for a slot at 10am tomorrow. Accept the offer. Hang up. Now check your calendar. If the booking has appeared with the address, contact details and notes — pass. If you got an SMS that says "you have a new enquiry, please call back" — fail. The whole point of an AI receptionist over voicemail is the booking, not the message.

Test 4 — The emergency test

Mention "burst pipe" or "no heat" or "gas leak" or "lockout". The AI should either (a) transfer the call to your mobile immediately, or (b) send you an instant SMS and book the next available slot. If it does neither, you've got a system that'll book emergency work into next Tuesday — which loses you the urgent-callout premium.

Test 5 — The discipline test

Ask the AI for prices it shouldn't quote — "How much would you charge me to install a kitchen?". A well-tuned system will say it needs to capture details and have you call back with a quote. A poorly-tuned one will hallucinate a number, which becomes a real problem when the customer holds you to it. The best systems treat custom-quote requests as a separate workflow, not a callable price.

Pricing reality: what you'll actually pay

Headline prices lie because they don't tell you about overage. Here's the actual cost at three realistic volumes:

A few patterns fall out, though always confirm current pricing on each provider's own page before deciding:

• Pay-as-you-go is meaningfully cheaper at low volumes. A solo trade taking 30-40 inbound calls a month (most lasting 3-5 minutes) sits well under any flat-rate threshold. At 120 minutes Invox is roughly £13 vs £99 on ARROW's flat plan.
• At 150 included minutes, ARROW's flat rate becomes the relevant comparison point. If your real volume sits at or above ARROW's bundle, the gap narrows and other factors (CRM, calendar booking, integrations) start to matter more than headline price.
• Above ~400 minutes a month, you'll need to ask each provider for an actual quote. ARROW directs higher-volume customers to a custom-quoted tier rather than publishing overage rates, so any number we put in this table for that volume on ARROW would be a guess.
• Hidden costs to watch for: per-call surcharges (some providers charge per pickup, not per minute), recording fees, transfer fees (some charge to bridge the call to your mobile), and "premium voice" upcharges. Read the small print before you sign.

The real question isn't "which is cheapest" — it's "which costs the least at the volume I'll actually use." Run the maths against your real call data.

What to do this week

The trades who get this right don't agonise. They run a small, fast experiment with real money on the line for 14 days and make the call.

1. Shortlist two. Based on the table above, pick the one trades-targeted AI option (ARROW, CallChimps, TradeMore or Invox) that fits your call volume, plus one alternative for comparison. Resist the urge to test five — you'll never get round to it.
2. Start free trials on both. Most UK options offer 7-14 day trials, often without a card.
3. Run the 5-test method. Twenty minutes total, both services. Score them honestly.
4. Forward your real number to whichever scored higher for one full working day. Use real customers as the judge — not synthetic test calls.
5. Decide. Look at calls handled, jobs actually booked into the calendar, and how it felt to use. Pick the one that wins, sign up, and start saving the £334-a-month that voicemail is currently siphoning out of your business.

If the answer ends up being Invox, great — we'd be glad to have you. If it's ARROW or CallChimps or TradeMore, we'll respect the decision a lot more than someone who guessed without testing. The point of this guide isn't to sell — it's to give you enough structure that you can make the call without us.