AI Business Plays

All plays › Local trades and home services

Local trades and home servicesClaude Code

Build the industry software you’re overpaying for with Claude Code, even if you don’t code

Picked and checked by Olga Pearson, host of The Practical AI Show · Published

Result

self-reported

A non-coding CEO built his own kitchen design software with Claude Code over 4.5 months. The tool his industry uses costs $6,600 up front plus $1,850 a year per seat. On air he designed a whole kitchen in about five minutes.

Randy Lorensen runs Premium Cabinets (30-plus locations). He’d never opened Claude Code, but he built TurnKey Cabinet Design alone on two Claude Max plans. A five-minute prompt on a Sunday night ran for three days and came back with a working version. He says the edge cases after that were 99% of the work. The software lays out cabinets using real manufacturer SKUs, with pricing and a parts list, and a bug button sends his spoken reports straight to the coding agents. He showed it on The Practical AI Show, Ep 61.

Who did it

Randy LorensenPremium Cabinets (30-plus locations)The Practical AI Show, Ep 61 (with Randy Lorensen) · Oct 16, 2026

Try it now

  1. Add up what your industry software costs you: up-front price, yearly fee per seat, training days and the hardware it needs.
  2. Write a plain-English spec of the result you want, e.g. ‘someone who has never designed a kitchen can do it without mistakes.’ Describe the result, not the code.
  3. Give that spec to Claude Code, let it build a first version, and expect the edge cases to take most of the time.
  4. Add a ‘report a bug’ button so you and your team can send problems straight back to the build, and test it in-house before you sell it.
Copy-paste prompt
I run [business] and we pay [$X up front + $Y/year per seat] for [software] to do [job]. I don’t code. Help me write a one-page spec for a simpler tool that lets [who] do [job] without mistakes. List the screens, the questions the tool asks, the data it needs (e.g. our product catalog with SKUs and prices) and what it outputs (e.g. a layout plus a parts list). Then list the 10 edge cases most likely to break it. Don’t write code yet.

Make it run every week

A play works once. Turn it into a routine in three small steps:

  1. Run it once and check the result yourself.
  2. Give your AI a schedule. Tell it when to run this play (every week, or every month) and to message you with what it found before you act.
  3. Keep a log. One line per run: what was checked, anything flagged, what you did.

See how owners turn plays into systems →

What to watch out for

Told on air. It’s a 4.5-month build with Claude agents running around the clock, so it isn’t a weekend project, and TurnKey is still in beta. Disclosure: Practical AI is hosted by Olga Pearson (who runs AI Business Plays) and Chris Pearson (who builds PageMotor). Build it with Page: a product page and waitlist for the tool you build.

Did it work?

PageMotor keeps this site free.

Related plays

  • Local trades and home servicesGrok Bot

    Have a Grok Bot negotiate your credit card processing fees

    self-reported

    About $5k a year saved on card processing for an auto repair shop that takes about $2M a year in cards. The markup over interchange went from 0.35% to 0.10%.

    Franklynn Automotive is an auto repair shop that runs about $2M a year through card payments. Processing is the bill you set up once and never look at again. Dan Mallatt gave his Grok Bot 24 months of statements. Interchange isn’t negotiable, but the processor’s markup is, so the bot found what the shop was paying on top of interchange and built the case for a lower rate based on the shop’s higher volume. It drafted every email and the counteroffer, and Dan reviewed each one before it went out. The markup dropped from 0.35% to 0.10% over interchange. His last step: get the final pricing in writing.

    By Dan Mallatt (@dan_mallatt), Franklynn Automotive (auto repair shop) · Source

  • Local trades and home servicesGrok Bot

    Have a Grok Bot build the moving inventory from customer photos

    self-reported

    A customer booked a whole move through the company’s Grok Bot, from quote to confirmed date, without a single phone call. The bot builds the full inventory list from photos or a video, and a person checks every line.

    Next Gen Movers is a long-distance moving company, and they say most moving quotes go wrong before moving day because nobody really knows what’s being moved. Now customers send photos or a quick video, and their Grok Bot goes room by room and builds the full inventory list. A real person on the team checks every line and asks about anything unclear. They also ask for photos of closets, stairs and where the truck will park, because that’s where the extra hours come from. One customer booked the whole move through the bot, quote to confirmed date, with no phone call.

    By Next Gen Movers (@nextgenmove), Next Gen Movers (long-distance moving company) · Source

  • Local trades and home servicesGrok Bot

    Have a Grok Bot build your landscaping website instead of paying a $2,500 quote

    self-reported

    His Grok Bot built a website for his landscaping business, found the cheapest domain and set everything up from the info and pictures he gave it. He had been quoted up to $2,500 for the same job.

    Max runs a landscaping business and needed a website. He had quotes as high as $2,500. Instead he gave his Grok Bot some details about the business and photos of his work. The bot built the site, compared domain prices, picked the cheapest one and set it all up. He says the result is exactly what he wanted.

    By Max (@max_peloquin1), Landscaping business · Source