Technology AI

How trade firms are using AI

What official surveys show about AI use in construction and small firms, the uses that save a heating firm time today, and where human checks must stay.

Ava research teamPublished 6 min read

On this page
  1. Where the trade stands
  2. What firms use it for
  3. Uses that save time today
  4. What holds firms back
  5. Keep a person in the loop
  6. Few firms plan to start soon
  7. Data and privacy basics
  8. Signs a tool is worth trialling
  9. Your customers already use AI
  10. Trying one use safely
  11. Worked example: the quote letter
  12. Frequently asked questions
  13. References

Key takeaways

  • Most construction firms still use no AI, but the share that does is growing.
  • Firms that use it mostly aim at running the business more efficiently, not at replacing customer contact.
  • Lack of expertise and difficulty fitting tools into existing systems hold firms back more than cost.
  • The practical uses today are writing, summarising, answering routine calls and triaging enquiries, all with human checks.
  • AI should never give gas safety advice, diagnose faults remotely or set prices on its own.

Where the trade stands

Official surveys give a clear picture. In late September to early October 2026, 81% of UK construction businesses said they did not use any AI technology[1]. That is still a large majority, but it has fallen from 91.9% in September 2023[1]. Across all businesses the share not using AI was lower, at 64.1%[1].

The gap is clearest for the tools most people mean when they say AI: text generation with large language models. 6.7% of construction businesses used it in autumn 2026, against 19% of all businesses and 18.8% of the smallest firms with up to nine employees[1]. Most plumbing and heating firms fall in that smallest group, so on these figures the trade is behind businesses of its own size.

Businesses using large language model text generation, autumn 2026
Businesses using large language model text generation, autumn 2026Column chart. Construction: 6.7%; Businesses with 0 to 9 employees: 18.8%; All businesses: 19%.0%5%10%15%20%Construction: 6.7%6.7%ConstructionBusinesses with 0 to 9 employees: 18.8%18.8%Businesses with 0 to 9employeesAll businesses: 19%19%All businessesBusinesses using large language model text generation, autumn 2026Column chart. Construction: 6.7%; Businesses with 0 to 9 employees: 18.8%; All businesses: 19%.ConstructionConstruction: 6.7%6.7%Businesses with 0 to 9 employeesBusinesses with 0 to 9 employees: 18.8%18.8%All businessesAll businesses: 19%19%

Source: Office for National Statistics, 2026[1]

Show the data
ItemValue
Construction6.7%
Businesses with 0 to 9 employees18.8%
All businesses19%

A separate government survey points the same way. Among SME employers that use any technology or web-based software, 34% used AI in 2025, including 32% of micro firms, making it the fourth most common software after booking software, CRM and HR systems[2].

What firms use it for

Construction firms that do use AI mostly point it at their own operations. In autumn 2026, 65.1% of construction businesses using AI used it to improve business operations, against 15.6% using it to provide or personalise services for customers[1]. Across all sectors, the most common areas for AI use or plans in 2025 were marketing and administration, each at 72%, followed by IT at 64%[3], and 65% of current and prospective users said the main reason was efficiency or productivity[3].

That matches what owners say they want: fewer evenings on paperwork, fewer missed calls and less retyping, rather than a robot talking to every customer.

Uses that save time today

Practical AI uses for a plumbing and heating firm
UseWhat it doesHuman check needed
Writing quotes and lettersTurns bullet notes into a clear quote cover letter or customer letterRead every word; check prices and scope
Summarising job notesTurns an engineer's voice note into a tidy job recordEngineer confirms before it is saved
Answering routine callsAnswers out of hours or when the line is busy, takes down the job and books routine visitsRules agreed in advance; urgent calls follow your rota
Triaging enquiriesSorts website and email enquiries by type and urgencyOffice reviews the urgent pile first
Review replies and postsDrafts replies to reviews and social posts from your own factsEdit for accuracy and tone
Explaining jargonRewrites technical notes in plain English for customersCheck nothing technical is lost
Searching manualsFinds the right section of long installation manualsAlways confirm against the manufacturer's document

Phone answering is where AI is most visible to customers. A basic AI receptionist takes a message; an AI Employee such as Ava takes down the job, follows your on-call rota for urgent calls and books routine visits into a tested diary. Whichever you consider, judge it on what happens after the call, not on how natural the voice sounds.

Notice what is missing from the list: diagnosing faults, deciding which engineer goes, approving work and setting prices. Those decisions need a qualified person who can see the system or knows the customer, and handing them to software creates risk without saving much time.

What holds firms back

Cost is not the main barrier. In autumn 2026, 31.5% of construction businesses had not tried to use AI, and among the rest the most-cited barrier was lack of AI expertise at 8.2%, ahead of cost at 5.9%[1]. Asked what would support AI use, 19% of construction businesses named ease of integration and use, and 10.2% named affordability[1].

In plain terms: owners do not know where to start and do not want another tool that does not talk to the job software. Both are fair concerns. The answer is to start with one narrow use, measure it, and only then add more.

Keep a person in the loop

Businesses that use AI mostly check its work. 84% of UK businesses using AI reported at least some human input or checking of AI outputs, and 67% reported significant checking[3]. In a heating firm, some outputs must always be checked, and some tasks should never be handed over.

  • Never: gas safety advice beyond your approved script, remote diagnosis presented as certain, or prices set without your rules.
  • Always checked: quotes, customer letters, anything sent in your name, and anything involving money.
  • Spot-checked: call summaries, enquiry sorting and job note summaries, reviewed daily at first and weekly once trusted.
  • Logged: what the tool did and who checked it, so mistakes can be traced and fixed.

A gas smell call is never a job for improvisation, human or AI. Whoever answers must give your approved safety message and the national gas emergency number, and flag the call as urgent.

Few firms plan to start soon

Plans are cautious too. In autumn 2026, 80.3% of construction businesses had no plans to adopt AI technologies in the next three months, and 11.9% were not sure[1]. For an owner who is curious, that has two implications. There is no rush to adopt something half-understood because competitors are doing it; most are not. But the firms that do find one or two reliable uses early will have them bedded in before customers start to expect them.

A sensible middle path is to watch, try one low-risk use, and keep notes on what works. That costs little and builds the expertise that the surveys say is the main thing missing.

Data and privacy basics

  • Do not paste customer names, addresses, phone numbers or photos into general AI tools unless you have checked how the tool stores and uses data.
  • Use business accounts for AI tools, not personal ones, so the firm controls the data and access.
  • If a tool handles customer data for you, it is a processor and needs a proper contract with the terms data protection law requires.
  • Tell customers when they are dealing with an automated system, such as on the phone, and give them a route to a person.
  • Keep recordings and transcripts only as long as you need them, and say how long in your privacy notice.

The guide on AI call handling and UK GDPR covers these points for phone tools in detail.

Signs a tool is worth trialling

  • It solves one problem you can name and measure, such as calls missed after five o'clock.
  • It works with the systems you already use, or the hand-off is a simple email or summary.
  • The supplier can explain clearly what it does, what it does not do and what happens when it is unsure.
  • You can trial it on real work for long enough to include a busy week.
  • You can switch it off and take your data with you.

Your customers already use AI

Customers are not strangers to the technology. In 2024, 96% of UK adults said they were aware of the term AI, and six in ten reported using AI chatbots in the past three months[4]. That does not mean every caller wants to speak to one, and some will not. It does mean that a clear, honest AI system that says what it is and passes calls to a person when needed is unlikely to surprise most people.

Trying one use safely

  1. Pick one pain point

    Missed calls, quote letters or job notes. Not all three.

  2. Write the rules

    What the tool may do, what it must never do and when it hands over to a person.

  3. Check the data handling

    Where customer data goes, who processes it, how long it is kept and what the contract says.

  4. Trial for a month

    Measure the time saved and every error, and read the outputs yourself.

  5. Decide

    Keep, adjust or stop. Write down why.

Worked example: the quote letter

Example: a fictional owner spends two evenings a week turning survey notes into boiler quotes. She starts using an AI writing tool for the cover letter only, feeding it her notes on the property, the recommended boiler and the customer's priorities, without names or addresses. She keeps prices and the scope of work in her job software's quote template, untouched by AI. Every letter is read and edited before it goes. After a month she has saved about an evening a week, and she has caught two letters where the tool described a feature the boiler does not have, which confirms the checking step must stay.

Frequently asked questions

Are plumbers using AI?

A minority are. Most UK construction businesses still use no AI, but use is growing, mostly for admin, writing and call handling rather than technical work.

What is the most useful AI tool for a heating firm?

It depends on your biggest time drain. For many firms it is handling calls when nobody can answer, or turning notes into quotes and job records, always with checks.

Is AI expensive for a small firm?

Cost is not the main barrier in official surveys; lack of expertise and integration with existing systems are cited more often. Start with one narrow use and measure it.

Can AI give customers technical advice?

It should not give gas safety advice beyond your approved script or diagnose faults as if certain. Keep technical judgement with qualified engineers.

References

  1. Office for National Statistics (2026). Business insights and impact on the UK economy: wave 165 dataset (AI tables) (opens in a new tab) ↑

  2. Department for Business and Trade (GOV.UK) (2025). Longitudinal Small Business Survey 2025: SME employers (businesses with 1 to 249 employees) (opens in a new tab) ↑

  3. Department for Science, Innovation and Technology (GOV.UK) (2025). AI adoption survey report 2025 (opens in a new tab) ↑

  4. Department for Science, Innovation and Technology (GOV.UK) (2024). Public attitudes to data and AI: Tracker survey (Wave 4) report (opens in a new tab) ↑

Book a demo

See Ava handle your calls.

Book a short demo. We’ll walk through your calls and workflow, and agree what Ava should take care of.

Book a demo