AI in home services, trade by trade

Home services is one of the fields where AI is doing real, boring, useful work right now. Not visions of robot dispatchers. Not chatbots booking your customers. Actual time savings on the parts of the job that owners hate doing. We talk to crew owners across plumbing, HVAC, electrical, landscaping, and cleaning every month, and the patterns shake out differently for each.

Most of the home-services owners we work with don't have time to read about AI. They want to know two things: where it actually helps, and where it's still a waste. We'll go trade by trade.

One thing is true across all of them, so we'll get it out of the way first. The single biggest time-saver in home services today is voice-dictated note cleanup. Tech finishes a job, opens voice mode on Claude or ChatGPT, talks through what they did. AI cleans it into something the customer can actually read. About twenty minutes saved per day per tech, on a $20/month phone subscription. If you're not doing this yet, start there before anything else on this page.

OK. Trade by trade.

Plumbers

Plumbing is the easiest fit for AI right now because so much of the job is admin. The pipe work is what you trained for. The paperwork is what eats your evenings.

Things plumbers we know are doing today:

• Voice-dictated estimates on the drive home from the job. Pricing is in your head; the writing-it-down is what costs you time. Claude's voice mode handles this well.
• Code lookups in plain English. "Walk me through the venting requirements for a basement bathroom in Massachusetts" gets you closer than flipping through the IPC.
• Customer-facing job summaries that don't read like a robot wrote them. Feed your dictated notes plus a short voice doc that explains how you talk to customers. The write-up comes back sounding like you.
• Permit-form starter drafts.

HVAC

If you've got three years of service history per customer, you're sitting on a gold mine. The trick is making it actually useful at the moment a tech walks into the house.

What we've seen work in the field:

• Service-history summaries on the way to the call. Pull the last three visits, AI summarizes the open issues and what was replaced last, the tech walks in ready instead of catching up.
• Follow-up notes that mention the specific work, not "thanks for your business." Two-second variants, generated straight off the job notes.
• Annual maintenance plan letters personalized for fifty customers in one prompt.

Electricians

This is the most safety-critical trade on the list, which sets a higher bar for what AI should touch and what it shouldn't.

Useful today:

• Permit paperwork drafts.
• Code questions framed as conversations. "I have a 200-amp main and a sub-panel for an ADU; what's the minimum feeder wire size?" gives you a fast first read. Verify before you stake your license on it.
• Bid letters for larger residential and small commercial jobs. The estimating is yours. The write-up is AI's.

Landscaping

Landscaping is an unusual case. Both the visual side and the text side of the job benefit, which most other trades can't say.

What's working:

• Photo-based starter quotes. Customer texts a photo of their yard; you drop it into Claude or ChatGPT with your standard pricing notes; you get back a draft quote in two minutes. Edit and send.
• Seasonal scheduling notes for clients. "Heads up, we're moving your fall cleanup to next Tuesday because of the storm" times forty addresses, drafted in one prompt.
• Crew schedule shuffles when weather kills a week. You tell the AI what got moved; it drafts the customer notes; you skim and send.
• Bid letters for larger residential and commercial jobs, same as electricians.

Cleaning crews

Cleaning is where customer trust does most of the selling. AI helps when it makes that trust more visible. It hurts you when it tries to fake it.

What's working:

• Pre-walkthrough write-ups. After a new-customer walkthrough, dictate what you saw, get back a clean estimate that lists exactly what's in scope and what's not. Customers love the specificity.
• Crew schedule reshuffles. Someone calls out sick at 6am; AI drafts the customer notes about a one-hour delay; you skim and send.
• Customer-specific cleaning checklists. Some homeowners want the dog hair on the stairs handled first. Some don't want anyone in the basement. Write the preferences down once, and your AI remembers them on every visit.

If you do one thing this month

Start with voice-dictated job notes. It's the lowest-friction project in any of these trades, and the savings show up in week one.

A few practical pointers before you start:

• Use ChatGPT or Claude. Both have voice modes that work in the truck.
• Pay for the better tier. The free version's voice mode is fine for a personal commute. For job notes, the paid version is worth the twenty dollars a month.
• Write one short voice document that explains how you talk to customers, and paste it at the top of your prompt. Without that, your output reads like everyone else's.
• Don't try to perfect the prompt before you start. The first week of using it badly is what teaches you what to fix in week two.

For a deeper read on why that one voice document changes everything, see our primer on context. For terms you might bump into reading the rest of the AI internet, the glossary is here.