AI for trades and field services

AI helps trades and field service businesses win more work and lose fewer evenings by taking over the office jobs: drafting quotes from a voice note, replying to enquiries and booking appointments while you are on the tools, and chasing unanswered quotes and unpaid invoices automatically. You do not need new software, a website rebuild or any technical skill to start. For most plumbers, electricians and builders the fastest returns come from three places, in this order: quoting, scheduling and follow-up. Each one can be automated in a small, contained way without touching how you actually do the work on site.

"Quoting at the kitchen table at 9pm is not a personality trait, it is a process failure. And the fix costs less than one lost job."

Dean Cookson, founder, Operosus

Why are trades businesses such a good fit for AI?

Because the bottleneck in most trades firms is not the skilled work. It is everything wrapped around it.

A Salesforce survey of 350 technicians and tradespeople found they estimate losing more than seven hours a week, close to a full working day, to low-value tasks such as manual data entry and writing up jobs. The same research found 47% of appointments do not go to schedule, and 81% of respondents believe AI agents could help them do their jobs more efficiently. The survey was run in the United States, but the shape of the problem, quoting at the kitchen table at 9pm and a diary held together by texts, will be familiar to any UK tradesperson.

The labour side makes this sharper in the UK. The CIOB and FMB State of Trade survey for the first half of 2025 found 61% of small and medium building firms were directly affected by shortages of skilled tradespeople, with 49% seeing job starts delayed and 23% cancelling jobs because of staffing gaps. When you cannot hire your way out, the only lever left is making the people you have, including yourself, spend more of the week on billable work.

And adoption is no longer a fringe activity. The Office for National Statistics reported that 23% of UK businesses were using some form of AI by late September 2025, up from 9% when the question was first asked in September 2023. The firms quoting faster and replying to enquiries within minutes are increasingly your direct competitors, not tech companies. The full adoption picture is in our UK small business AI statistics table.

As Taksina Eammano, EVP and General Manager of Field Service at Salesforce, put it in that research: "To maintain critical infrastructure and the level of service today's consumers expect, AI is a necessity."

What can AI take off your plate first?

Start with the jobs that are repetitive, written down and follow a pattern. That is exactly what current AI tools are good at.

None of these touch the work itself. They compress the admin wrapped around it.

How does AI speed up quoting?

Quoting is usually the best first project because it is high value, high friction and entirely text-based.

The pattern that works looks like this:

1. You record a voice note on the drive back: rough scope, materials, anything unusual about the job.
2. An AI tool drafts the quote using your price list, your standard wording and examples of quotes you have sent before.
3. You read it, adjust anything that needs judgement, and send it. The human check stays in; only the typing goes.

The same pattern scales up to formal bids. At Operosus we built Bidwell, a tool that drafts tender responses for small firms from their own past bids, policies and evidence. The lesson from building it applies to any trades business: the draft is only as good as the source material you give it. A folder of your ten best past quotes, your current price list and your standard terms is worth more than any clever prompt. Assemble that first and almost any decent AI tool can produce a usable first draft.

What AI should not do is set your prices. Margins, contingency and gut feel about a difficult customer stay with you. Treat the AI as a fast typist that knows your paperwork, not as an estimator.

Can AI answer enquiries and book jobs while you are on the tools?

Yes, and for most trades firms this is where the most money is leaking. An enquiry that waits until the evening for a reply is often an enquiry that has already gone to whoever answered first.

The reliable pattern is a connected chain rather than a chatbot bolted onto a website:

• The enquiry arrives through a simple form, WhatsApp or a missed-call text-back.
• An automated reply goes out within seconds, asking the two or three questions you would ask anyway: postcode, type of job, photos if useful.
• Qualified enquiries are offered real appointment slots from your calendar and book themselves in.
• Confirmations and reminders go out automatically, which is what actually cuts no-shows.

We have built this end-to-end for a mobile veterinary service, where families book home visits through the website and the system handles confirmation, payment links and the follow-up communications around the appointment without anyone touching a keyboard. A mobile vet and a mobile plumber are operationally the same business: a skilled person, a van, a diary and customers who want certainty about when someone is turning up. The same flow transfers directly.

What about scheduling across a team?

Once there is more than one person on the road, the problem shifts from booking to allocation: who is nearest, who has the right skills, who has slack tomorrow. This is a matching problem, and it is one AI-backed systems handle well. We saw the same shape of problem building a venue-hire platform for schools, where the job was matching demand in an area to nearby capacity. For a trades firm the equivalent is routing the right person to the right job with the least dead driving time. You do not need this on day one, but it is where scheduling automation goes next once basic booking is solved.

How do you stop quotes and invoices disappearing into silence?

Follow-up is the cheapest revenue in the business, because the hard work, getting the enquiry and pricing the job, is already done. It fails in most firms for one reason: nobody enjoys chasing.

An automated follow-up sequence is simple and unglamorous:

• Day 3 after the quote: a short message checking it arrived and asking if there are questions.
• Day 7: a nudge with a gentle deadline, for example how long the price holds.
• Day 14: a final note leaving the door open, then the lead is marked closed.

Invoices get the same treatment on the other side of the job, with polite, consistent reminders that do not depend on your mood or your memory. Our guide to AI invoice chasing covers the full reminder ladder. AI earns its keep here by writing messages that sound like you rather than a robo-letter, and by reading replies: a "yes go ahead" gets booked in, a "we've gone elsewhere" closes the record, and anything ambiguous gets flagged for a human.

Add one more message to the end of the chain: the review request, sent a day or two after the job closes while the customer is still pleased. For a local trade, a steady stream of recent Google reviews does more for enquiries than most paid advertising.

Do you need to replace your existing software?

No, and be wary of anyone who says otherwise. The best results come from connecting what you already have, even if what you already have is a calendar, a spreadsheet and WhatsApp.

A sensible rule of thumb:

• Keep: your accounting tool, your calendar, your phone number, anything customers already use to reach you.
• Connect: enquiry sources into one place, your diary into the booking flow, your job records into invoicing.
• Add: the AI layer that drafts, replies, books and chases on top of those connections.

Most of the booking and follow-up patterns above are built from off-the-shelf parts. A small automation platform, a form, a calendar connection and an AI model for the reading and writing steps. The craft is in the wiring and the wording, not in exotic technology.

What should you avoid?

• Letting AI send prices unchecked. Drafts yes, autonomous quoting no. A wrong number in writing is a commitment.
• Pasting customer details into free consumer AI tools. Names, addresses and job details are personal data. Use business-grade tools with proper data terms, and check where the data goes.
• Chatbots that bluff. An assistant that invents an answer about gas safety or wiring regulations is a liability. Constrain automated replies to scheduling and information-gathering, and route technical questions to a human.
• Big-bang projects. A six-month "digital transformation" is how trades firms end up with shelfware. One workflow, live, in use, then the next.
• Anything that adds steps for the customer. If booking now takes longer than a phone call did, the automation has failed regardless of how clever it is.

Where to start

Pick the single workflow that costs you the most evenings, which for most trades businesses is quoting, and automate only that.

1. Gather your raw material. Your ten best quotes, your price list, your standard terms. One folder.
2. Run two weeks of drafts. Voice note in, AI draft out, you edit and send. Measure nothing more sophisticated than how quotes went out same-day versus before.
3. Then move to response and booking. Instant replies to enquiries, real diary slots, automatic confirmations and reminders.
4. Then switch on follow-up. The three-message quote chase and the post-job review request.

Each step pays for the next one, and none of them requires changing how you do the work that customers actually pay for.

If you would rather have it built than build it, this is what Operosus does: small, connected AI systems for UK businesses, wired into the tools you already use. Our trades industry page shows the specific systems we build for firms like yours, and our AI appointment booking guide goes deeper on the diary side. Get in touch and tell us which evening job you want rid of first.