The first message came in on a Sunday night. "We're losing patients on Tuesdays and I cannot figure out why." Three sentences. No data. Just the gut feeling of a dentist who'd been watching his own books for nine years and knew when something was wrong.
Monday morning I was at a corner table in Crate Cafe with a cold brew and a fresh terminal window. I asked for read-only access to their patient inquiry log. They sent it within an hour. Eight months of CSV. Every inquiry, every booking, every cancellation, every reschedule.
By the second coffee I could see it. Tuesday inquiries had a 41% no-show rate. Every other day hovered around 12%. The data wasn't subtle — it was screaming.
Tuesday was the day their reception staff was lightest. Two people instead of four. The follow-up call that would normally happen 24 hours before the appointment was getting deprioritized. Patients who'd inquired Tuesday weren't getting the gentle reminder nudge that Wednesday-through-Friday patients got. They were drifting.
Wednesday I built the fix in a single n8n flow. Inquiry comes in → tag with day-of-week → if Tuesday, route through a GoHighLevel SMS sequence that sends a personal-feeling reminder at the 48-hour, 24-hour, and 3-hour marks. The SMS copy was written in the dentist's own voice — I had him record a 90-second Loom describing how he'd talk to a hesitant patient. I transcribed it and used it as the prompt.
Thursday the first Tuesday-cohort patients started arriving. Show rate moved from 59% to 84% in the first three-week window. By the end of the month the practice was running net positive on Tuesdays for the first time since the previous summer.
The thing nobody tells you about automation is that the hardest part isn't the workflow. The hardest part is finding the loop. You have to sit with the data long enough to see what's actually happening — and you can't do that with eleven Slack channels open. You have to close the laptop on everything else and just look.
