Click and Close

Click and Close sells an AI receptionist to businesses that lose leads in the gap between someone filling in a form and someone calling them back. I built both halves of it: the receptionist itself, and Lighthouse, the control centre clients use to run it.

The receptionist is assembled from n8n, Retell AI and the Google Calendar API, and it configures itself per client. Onboarding is basic information in, working receptionist out — no bespoke engineering for each account, which is what makes the product sellable rather than a consulting engagement.

It answers inbound calls, but its real job is outbound. When someone submits a form on a client's site, it calls them back immediately — and that immediacy is the whole product. Callers feel attended to instead of queued, and our clients convert more of them. On the call it books the appointment, writes it to both the caller's and the client's calendar, and emails the client a summary of what was said.

Lighthouse is where clients manage all of it. The home dashboard tracks call volume, how many calls convert into appointments, caller satisfaction, and average call duration — plus a form that rings their own phone when they need us. A calendar page shows every appointment booked. Call logs pair each recording with a summary, the caller's context from before the call, and call metadata, so clients can audit quality themselves. Settings let them shape what the receptionist knows about their company and what it should aim for on an inbound call.

I still own both products — maintaining them, fixing what breaks, and building against what clients ask for next.