All plays › Local trades and home services

Local trades and home servicesClaude (Sonnet)

505 client conversations reviewed for about $60 to find why a cleaning business stopped growing

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

What was the result?

self-reported

Roy Banwell says Claude pulled and analyzed the conversations with all 505 new clients his cleaning business sold in 6 months, for about $60 in Sonnet tokens. It showed fewer than 60% said the service went well, another 15 to 20% said nothing, and the rest felt they got bad service. Based on that he decided to cut ad and sales spend by over 50%, which he expects to save over $4,000 a month, and fix service quality first.

Roy Banwell quit his job in Feb 2026 to run his wife's cleaning business. Sales were strong, 505 new clients in 6 months, yet the business was flat, and getting those sales cost over $10,000 a month, close to 20% of gross revenue, in ads, agency and sales reps. Instead of guessing, he had Claude read every client conversation. The answer was a quality problem, not a sales problem. His plan: spend less on ads, improve hiring, scheduling and cleaners, and ramp sales back up only when more than 80% of clients are happy and fewer than 10% are unhappy.

Who did this?

Roy Banwell (@Roy_Banwell)Home cleaning business; Miami, FL (per his profile)Roy Banwell (@Roy_Banwell) on X · Nov 6, 2026

How do I try it?

  1. Editorial adaptation: export every conversation with new clients from a set period (texts, emails, chat).
  2. Have AI tag each client after their first service as happy, said nothing, or unhappy, with the main reason and a short quote.
  3. Ask for totals, percentages and the top complaint reasons, plus patterns by cleaner, service type or day.
  4. Pick one fix and set a target, such as more than 80% happy and fewer than 10% unhappy, before spending more on ads.
  5. Rerun the same check every month and compare.
Copy-paste prompt
Editorial adaptation, not Roy Banwell's exact prompt. Here are my conversations with every new client from the last [6 months]. For each client, tag how they felt after their first service: happy, said nothing, or unhappy, and give the main reason in a few words with a short quote. Then give me totals and percentages, the top 5 complaint reasons, and any patterns by cleaner, service type or day. Do not contact any client; this is for my review only.

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 should I watch out for?

Self-reported. The 'over $4,000 a month' is the saving he expects from cutting spend, not a result yet, and he expects revenue to drop 15 to 20%. AI sentiment tags can be wrong, so spot-check a sample yourself. Client messages hold personal data, so use a tool and plan you trust with it. Build it with Page: a short feedback page clients get after every clean.

Did it work for you?

PageMotor keeps this site free.