Key takeaways
- 17.1% of construction businesses used AI in a business function in late 2026, below the 24.2% rate for all businesses[1].
- Small-business use of generative AI has climbed quickly, from 23% in 2023 to 58% in 2025 in a U.S. Chamber survey[2].
- Most real uses in plumbing are narrow: answering and logging calls, drafting office writing, and summarizing information for a person to check.
- AI has not changed headcount at most construction firms that use it; 87.6% reported no change[1].
The short answer: fewer than you'd think, growing faster than before
Trade press makes it sound as if every contractor has an AI strategy. The government's own survey says otherwise. The Census Bureau's Business Trends and Outlook Survey (BTOS) asks firms every two weeks whether they used AI in any business function. In the two weeks to October 18, 2026, 17.1% of construction businesses said yes, compared with 24.2% of all US businesses[1]. Plumbing sits inside construction in that survey, so the best available estimate is that roughly one construction firm in six is using AI in its operations.
- Construction
- All businesses
Source: U.S. Census Bureau, 2026[1]
Show the data
| Item | Construction | All businesses |
|---|---|---|
| Used AI in a business function | 17.1% | 24.2% |
| Employees used generative AI for work | 19.8% | 26.8% |
The direction is clear, though. Census researchers tracking the same question found AI use across all businesses rose from 3.7% to 5.4% between September 2023 and February 2024[3]. The current all-business rate of 24.2%[4] is several times higher than that early reading, though survey questions and methods change over time, so treat the comparison as a direction rather than an exact multiple. Plumbing is not leading, but it is moving.
Different surveys measure different things. The U.S. Chamber of Commerce asked small businesses whether they use generative AI at all, including for writing and research, and found 58% did in 2025, up from 40% in 2024 and 23% in 2023[2]. That higher figure counts an owner drafting an email with a chatbot. The Census figure is stricter: it asks whether the business itself used AI in a business function during a recent two-week period.
Source: U.S. Chamber of Commerce Technology Engagement Center, 2025[2]
Show the data
| Period | Value |
|---|---|
| 2023 | 23% |
| 2024 | 40% |
| 2025 | 58% |
Where plumbing shops are using it
Look past vendor marketing and the real uses fall into a short list. Most share three traits: the task is repetitive, the output is checked by a person, and a mistake is cheap to catch.
| Area | Typical use today | Who checks the output | Risk if it goes wrong |
|---|---|---|---|
| Phones | Answering overflow and after-hours calls, taking job details, routing urgent calls | Dispatcher reads the job summary | Missed emergency or wrong address |
| Scheduling | Booking routine visits into set slots, moving existing appointments | Office reviews the next day's board | Double booking or wrong job type |
| Office writing | Drafting emails, job descriptions, policy documents | Office manager edits before sending | Wrong promise to a customer |
| Estimates | Turning tech notes into clear scope wording | Service manager approves price and scope | Scope gaps or invented claims |
| Marketing | Drafting review replies, social posts and service pages | Owner or marketing lead | Privacy slip or misleading claim |
| Bookkeeping | Categorizing transactions, summarizing reports | Bookkeeper or accountant | Misfiled expenses at tax time |
What is rare: AI that diagnoses a plumbing problem, prices a job without human review, or dispatches technicians on its own. Those jobs carry license, safety and liability weight, and most owners keep them with people.
Phones are the most common starting point
The phone is where AI meets a plumbing company's revenue most directly. Calls arrive at all hours, many are urgent, and a missed call often means a lost job. That makes call handling a natural first use, and also the one where testing matters most. A useful AI phone agent has to take down the address and problem accurately, recognize an emergency, reach the on-call technician and leave a clean record. Our buyer's guide to AI phone answering covers the tests to run before you commit.
Ava, for example, is built around that narrow job for plumbers: answering calls, capturing job details, booking routine work into a tested schedule and sending a summary with the recording. Missed-call text-back and field-service software connections are still Planned, not Live, which is the kind of distinction worth asking every vendor to make.
What AI is doing to jobs
The fear is that AI replaces office staff. The data so far shows little of that in construction. Among construction businesses, 87.6% said AI use had not changed their headcount over six months, 4.2% said it increased and 3.1% said it decreased[1]. The Chamber found 82% of small businesses using AI had increased their workforce over the past year[2], though that does not prove AI caused the hiring; growing firms may simply adopt more tools.
In practice, plumbing shops that adopt AI tend to move office time away from repetitive intake and toward work that needs judgment: calling back unhappy customers, chasing estimates, coordinating parts. Our guide to talking with your office team about AI covers how to frame that change.
Why construction trails other industries
- The core work is physical. No software unclogs a drain, so AI touches the support work around the job rather than the job itself.
- Firms are small. Most plumbing businesses have a handful of staff, and nobody's full-time role is evaluating software.
- Field software already absorbs attention. Many shops are still finishing a move from paper or spreadsheets to a field-service platform.
- Mistakes are visible. A wrong address or a missed emergency is felt the same night, so owners are right to be cautious.
- Employees are using it informally. 19.8% of construction businesses said employees used generative AI for work tasks, against 26.8% of all businesses[1], often without any company policy.
What the next six months look like
The same Census survey asks whether firms expect to use AI in the next six months. Among construction businesses, 19.4% said they did[1], only a little above the 17.1% already using it. Across all businesses, 27.6% expected to use AI and 26.2% said they did not know[4]. That large group of undecided firms matters for plumbing owners: it means the tools are still being sorted out, prices and terms vary widely, and you do not need to rush.
Expect the change to come from inside software you already use more than from new standalone products. Field-service platforms, phone systems, accounting packages and email tools are adding AI features into their existing screens. That makes adoption easier, but it also means AI can be switched on in your systems without anyone deciding it should be. Check release notes and settings after each major update, and decide feature by feature what stays on.
The rules to write before anyone uses AI
Because staff are already experimenting, the first useful step is often a short written rule rather than a new tool. One page is enough if it covers these points:
- Which tools are approved for company work, and which accounts staff should use with them.
- What may never be pasted into an AI tool: customer names with addresses, card numbers, gate codes, passwords and employee records.
- Who reviews AI-written text before it goes to a customer, and that the reviewer owns what is sent.
- That AI must not quote prices, promise arrival times or describe code requirements unless a person has checked them.
- How to report a mistake, so problems are fixed rather than hidden.
Choosing your first use: a four-question test
Before buying anything, score each candidate task against four questions. A task that scores yes on all four is a good first pilot.
- Does it happen at least daily? Rare tasks never give you enough examples to judge quality.
- Can a person check every output in under a minute? A job summary is quick to read; a full estimate is not.
- Is the cost of a mistake small and recoverable? A badly worded email draft is; a missed burst-pipe call is not, so phone pilots need tighter rules.
- Can you measure it before and after? Answer rate, booked jobs and time spent are easy; customer goodwill is not.
Then run a time-boxed trial with clear stop rules, as described in our 30-day AI pilot plan.
Example: a nine-person shop's first year with AI
Example: a residential shop with six technicians and three office staff starts by writing a one-page rule on generative AI: no customer names, addresses or card numbers pasted into public tools. In the spring it trials AI call handling for evenings and weekends only, reviewing every call summary for two weeks before trusting the routing. In the summer the service manager starts using AI to turn tech notes into cleaner estimate wording, always approving the final scope and price. By the end of the year, nobody's job has gone; the office has stopped taking messages by hand at night and spends more time following up open estimates.
A good first AI use is boring, frequent and easy to check. Save the ambitious ideas for after you have measured the first one.
Frequently asked questions
What share of plumbing companies use AI?
There is no plumbing-only figure from the government. Census data for the construction sector, which includes plumbing contractors, showed 17.1% using AI in a business function in October 2026.
What is the most common AI use for plumbers?
Handling phone calls, especially overflow and after-hours calls, followed by drafting office writing such as emails, estimate wording and review replies.
Will AI replace my dispatcher or CSR?
Survey data shows most construction firms using AI have not changed headcount. AI usually takes the repetitive intake work, while people keep judgment calls, escalations and customer recovery.
Is it safe for staff to use free AI chat tools?
Only with rules. Write down what may never be pasted in, such as customer personal details, card numbers and passwords, and check every output before it reaches a customer.
References
U.S. Census Bureau (2026). Business Trends and Outlook Survey (BTOS), response estimates by sector (opens in a new tab) ↑
U.S. Chamber of Commerce Technology Engagement Center (2025). Empowering Small Business: The Impact of Technology on U.S. Small Business (2025) (opens in a new tab) ↑
U.S. Census Bureau, Center for Economic Studies (2024). Tracking Firm Use of AI in Real Time: A Snapshot from the Business Trends and Outlook Survey (CES-WP-24-16) (opens in a new tab) ↑
U.S. Census Bureau (2026). Business Trends and Outlook Survey (BTOS), national response estimates (opens in a new tab) ↑