Most chatbot use-case lists are written for every business at once, which means they land properly on none of them. Customer support, order tracking, product recommendations, cart recovery.

Genuinely useful if you run an online shop. Close to useless if you fit bathrooms, run a physiotherapy clinic or clean commercial kitchens.

Service businesses have a narrower problem, and a more solvable one. Someone finds you at 9pm, wants to know whether you cover their postcode, roughly what a job like theirs costs, and whether you've handled one before. There's nobody there to answer, so they open a new tab and ask the next company down the page.

Every use case below exists to close that specific gap. We've ordered them the way a real visitor moves through your site rather than by how impressive they sound, and each one comes with a short worked example so you can picture it running on your own pages.

What a service business chatbot actually is

Forget the button-tree bots of ten years ago, the ones that offered you four options and got sulky when you typed a fifth.

A modern site chatbot reads your own material. Service pages, price guides, coverage areas, past job photos, spec sheets, the answers you've typed into a hundred emails. It answers in your words, and when it doesn't know something it says so and takes a name instead of guessing.

That distinction matters for what follows. Several of these use cases only work because the bot can pull a real photo or a real PDF into the conversation, not just paste a link and hope.

1. Booking appointments without the phone tag

Phone tag is the tax busy service businesses pay for being busy. You call back, they're at work. They call back, you're under a sink.

A chatbot that's wired into your calendar removes the game entirely. It checks whether the job type is one you take, asks for a postcode, offers real availability, writes the booking in and sends the confirmation.

Worked example. A domestic electrician gets an evening visitor asking about EV charger installation. The bot confirms the service, checks the postcode falls inside the coverage area, offers three survey slots for the following week, takes a name and mobile number, and drops the appointment straight into the diary. The electrician sees it on the van in the morning.

The mechanics of connecting a calendar, handling cancellations and stopping double-bookings deserve more room than we can give them here, so if this is your priority use case, read how to let customers book appointments automatically next.

2. Giving an instant ballpark quote

This is the use case most service businesses talk themselves out of, and the one that moves the most money.

The objection is always the same. Every job is different, so we can't quote without seeing it. Fair enough. But the visitor isn't asking for a fixed price, they're asking whether you're in their universe at all. Silence sends them to whoever will give them a number.

A range, with an honest caveat attached, answers the real question and keeps them talking.

Worked example. A removals firm's bot asks four things. Bedrooms, floor level at both ends, distance, and whether packing is included. It comes back with "most moves like this land between £850 and £1,200, subject to a video survey", then offers to book the survey. Nobody is misled, and the enquiry survives.

The trick is in the questions and the guard rails: too few and the range is meaningless, too many and people abandon halfway. We've broken down how to build a quote flow visitors finish if you want the detail.

3. Answering the questions your FAQ page buries

Your FAQ page probably has thirty accordions on it. Almost nobody opens more than two.

The questions themselves are fine. The format is the problem, because a visitor with one specific worry has to hunt for it among twenty-nine they don't care about.

A chatbot inverts that. They ask, it answers, in one exchange. And it answers the awkward phrasings your page doesn't cover, the ones people actually type.

Worked example. A dental practice fields the same three questions all day. Do you see nervous patients, is there parking, do you offer payment plans. All three are on the FAQ page. None of them get read. The bot answers each in a sentence and, when the answer is yes, offers a consultation slot on the spot.

There's a second benefit here that's easy to miss. Every question the bot couldn't answer is a log of what your website fails to explain, which is the cheapest content research you'll ever get. More on that in our piece on answering FAQs without losing leads.

4. Capturing leads and qualifying them with the right questions

A contact form takes whatever the visitor decides to type. A conversation asks the follow-up question.

That difference is the whole use case. "Hi, can you help with a leak?" is not a lead. "Ground floor bathroom, tenanted property, landlord paying, needs someone this week, SW17" is a lead you can act on before you've spoken to anyone.

Good qualification also protects your time in the other direction. The enquiry that was never going to fit gets a courteous answer and a pointer elsewhere, without eating twenty minutes of yours.

Worked example. A commercial cleaning company asks four qualifying questions in the chat. Premises type, approximate floor area, frequency, and when they need to start. A 3,000 sq ft office wanting nightly cover goes to the sales inbox flagged as priority. A one-off domestic clean gets a polite no and a recommendation to search elsewhere. Neither one costs a phone call.

5. Catching after-hours and weekend enquiries

Domestic service enquiries cluster at the times people aren't at work. Evenings, Sunday mornings, the gap between putting the kids to bed and going to bed themselves.

Your office hours are 8 till 5. The mismatch is structural, and it's where most service businesses quietly lose their best leads.

A bot doesn't fix the fact that you're asleep. It fixes the fact that a visitor at 10:40pm currently gets nothing, has no reason to wait, and has four other tabs open.

Worked example. A boiler repair firm's site gets an enquiry at 22:40. The bot establishes there's no heating and no hot water, flags it as urgent, takes the address and a mobile number, books the first slot at 07:30 and tells the customer exactly when someone will arrive. The engineer reads the whole thread over breakfast.

Speed of response is not a soft benefit. It's one of the most consistently measured findings in B2B sales research.

6. Showing photos of past work inside the conversation

Service work is bought on proof. People want to see the thing you built for someone whose house looks like theirs.

Most sites handle this with a gallery page, which is a fine archive and a terrible answer. A visitor with a specific question ("what does resin bound look like on a sloped driveway?") is not going to scroll ninety images to find out.

A chatbot can pull the right two or three photos into the chat window as part of its answer. Question asked, proof supplied, conversation continues.

Worked example. A landscaping company gets asked about garden rooms on uneven ground. The bot returns three photos from previous jobs with a line of context under each, including one on a sloping plot, then asks whether the visitor would like a site visit. The visitor has effectively had a portfolio review without leaving the page.

7. Handing over brochures, spec sheets and PDFs on demand

Almost every service business has documents that do real selling. Fee schedules, service brochures, technical spec sheets, warranty terms, capability statements for tender submissions.

They usually sit behind a form on a downloads page, or worse, get emailed manually on request the next working day.

A chatbot can hand over the right document mid-conversation, and it can tell which one is right because it understood the question. That's a materially different experience from a downloads page listing eleven PDFs with similar names.

Worked example. A commercial flooring contractor is asked about slip resistance in a wet kitchen area. The bot explains the relevant rating in a sentence, delivers the spec sheet for the two suitable ranges, and asks whether they'd like a sample pack posted out. The email address gets collected at the sample stage, where it's genuinely earned.

Our own view on gating is unfashionable. Gate almost nothing. Let people read the brochure, and ask for details at the point where you're offering something that costs you money to provide, like a sample, a survey or a bespoke quote. The exception is high-value B2B material such as tender documentation, where knowing who's asking is part of the job.

Which two should you start with?

Seven use cases is a menu, not a to-do list. Trying to switch all of them on in week one is the reliable way to end up with a bot that does everything vaguely and nothing well.

Pick two. Almost always, the right pair is the question you personally answer most often on the phone, plus after-hours capture, because that one earns money while you sleep and takes almost no configuration.

• List the five questions you answer most on callsTake them from last week's actual calls, not memory
• Check which of those the bot could answer from existing pagesIf the answer isn't on your site, write it once and both benefit
• Decide your quote guard railsJob types you'll range-price, and the ones that always need a survey
• Tag ten past-job photosService, property type, notable feature
• Pick the one PDF worth handing over unprompted
• Agree what happens to an out-of-hours lead by 9amThe bot books it, a human still has to honour it

And a word on limits, because the honest version is more useful than the sales version.

That last point is the one people skip. A chatbot is not a set-and-forget appliance, it's a member of staff who learns from being corrected. Half an hour a month reading conversations will improve your conversion rate more than any amount of initial configuration.

If you'd rather watch these use cases running than read about them, you can see our chatbot in action on a live site and try to break it.