Book homeowner appointments overnight by texting with OpenAI and Zapier
A Minnesota appointment-setting firm reads homeowners' late-night text replies with OpenAI and books contractor visits automatically; it says 20 to 50 extra a day.
When a homeowner replies by text, a Zapier workflow uses OpenAI to pull out the date and time, checks contractor availability by zip code and service, and books the visit. Open-ended replies go to ChatGPT, which drafts a reply.
How it went
The company says 80 to 90% of first replies are handled automatically and that after-hours booking adds 20 to 50 appointments a day.
Worth knowing
The results are company-reported in a Zapier case study, including a $134M client revenue claim. It is a US firm and a high-volume setup built by a technical team. Texting customers in the UK needs consent under PECR, and bookings should be spot-checked.
Try it yourself
Paste into ChatGPT, Claude or another AI assistant and fill in the square brackets.
I take inquiries by text for [business type] in [area]. Write rules for an AI assistant that reads a customer's reply, pulls out the date and time they want, and proposes a slot from my availability: [availability]. List which replies must go to a person, such as complaints or unusual requests. Show me the rules and three sample replies before it contacts any customer.
Source: Zapier. Summarised in our own words; read the original for the full account.
The owner of a small Ottawa interior design firm uses ChatGPT for emails and construction scheduling, plus an AI plugin to show clients photorealistic renderings.