Case study: a 9-minute lead response time, built by one operator

Marcus runs a five-person commercial cleaning company. Inbound leads arrived by web form and email, sat until someone had time, and were answered in one to two business days. Industry data — and his own close rates — said speed mattered more than anything else he could fix.

The build

Following the Professional track modules in order, Marcus built: a lead classifier (Module 3) scoring inbound inquiries against his ideal customer profile and returning JSON with a tier and reasoning; an automation (Module 4) that triggers on form submission, runs the classifier, writes to his sheet, and alerts his phone for hot leads; and a follow-up generator (Module 7) drafting a personalized response from the inquiry details in his voice.

Total build time: two weekends. Tools: his existing form provider, Make, and one model API. No code beyond pasting a JSON schema.

The results and the honest caveats

Hot leads now get a personalized response in under ten minutes around the clock — drafted by AI, approved by Marcus from his phone (the human gate stayed; a wrong quote is expensive). Close rate on hot-tier leads rose meaningfully within the first quarter, though Marcus is careful to note that a busy season overlapped the rollout.

Two failures worth learning from: the classifier initially scored large enterprises as 'hot' although Marcus can't serve them — the ideal customer profile needed a size ceiling, not just a floor. And one week of silence from the alerting revealed the automation had failed quietly after a form field was renamed — which is why failure alerting (Module 4's '2 AM answer') became part of his spec. Both fixes are now permanent test cases.