The system segments patients who haven't visited in 18 months, drafts AI-personalized win-back messages that reference their last exam and eyewear purchase, and drops responders directly into open booking slots — turning a dormant patient database into a steady stream of filled appointments.
Typical results
24%
Win-back response rate
4x
Response vs. generic blasts
240
Reactivated bookings per month
Figures reflect typical results for this kind of automation.
The challenge
The chain's database held over 14,000 patients who hadn't visited in a year and a half — prescriptions aging, frames wearing out, no reason to come back. Generic 'we miss you' SMS blasts pulled a 6% response, front-desk staff had no time for call-downs, and marketing budgets chased new patients instead.
Before and after
Win-back response rate
4×Reactivated bookings per month
6×Midweek slot utilization
+42%The automation
We built a segmentation flow that pulls every patient at 18-plus months since their last visit, grouped by branch, age, and purchase history. OpenAI drafts each win-back message individually — referencing the last exam date, the lenses or frames they bought, and why a recheck matters now — and sends in paced WhatsApp batches. Responders see live open slots and book in two taps; every outcome writes back to the CRM for the next cycle.
How it works
- 01
Segment the dormant
Every month, patients 18-plus months since their last visit are pulled and grouped by branch.
- 02
AI drafts each message
OpenAI writes each win-back individually, referencing the last exam date and the eyewear purchased.
- 03
Paced sends
WhatsApp batches go out sized so each branch can actually absorb the responses.
- 04
Two-tap booking
Responders see live open slots and book without ever talking to the front desk.
- 05
Learn and repeat
Outcomes write back to the CRM; non-responders re-queue for a different angle next cycle.
Starts from
- Dormant patient files
- 18-month segmentation
Delivers
- Personalized WhatsApp
- Open-slot booking
- Response tracking
The impact
Over a typical eight-month run, about 1,900 lapsed patients rebook exams — response rates rise from 6% to 24% once messages reference each patient's own history. Reactivated patients fill quiet midweek slots, most update their eyewear at the visit, and the campaign runs monthly without anyone touching it.
Questions about this automation
How are lapsed patients selected each month?
Once a month the automation pulls every patient whose last visit was 18 or more months ago and groups them by branch, age and purchase history. Patients who already have an upcoming appointment or who opted out of messages are excluded, so the list only contains people who actually need a recall.
How are the win-back messages personalized?
OpenAI drafts each message individually from the patient's record, mentioning the date of the last exam and the lenses or frames bought at that visit, along with a short reason a recheck matters now. Messages follow approved templates and tone rules, and they are sent in paced WhatsApp batches.
How does the automation avoid flooding a branch with bookings?
Batch sizes are set per branch based on available appointment slots, so each location receives only as many messages as it can absorb. Responders see live open slots and book in two taps, and non-responders are queued for a different message angle in the next monthly cycle.