Incoming messages handled with judgment.
Never lose inquiries that arrive after closing time.
Updated August 2026·4 min read
The situation
A dental clinic closes at 19:00. The message arrives at 21:40: “how much is an implant? are you open on Saturdays?”. The next morning someone reads it at 9:30, answers at 11:00, and by then the person has already asked two other clinics.
This is not a customer-service problem, it is a scheduling one. The question arrives when nobody is there, and the value of answering drops with every hour that passes.
What makes the answer possible
The agent does not answer just because a message arrived. It answers because the campaign is active and carries what it needs to answer with.
- An active campaign. Answering outside business hours is not a global switch on the product: it happens inside campaigns the organization turned on.
- The documents. The current price list and the opening hours. They are the source of every fact the agent states — if it is not there, it does not say it.
- The 24-hour window. WhatsApp allows replies within 24 hours of the customer's last message. The window stays visible in the inbox for the whole team.
How it operates
At 21:40 the agent answers with the price list in hand: the initial assessment is free, and Saturdays run from 9:00 to 13:00. Two facts, both citable.
When the conversation moves onto clinical ground —“does it work if I lost a tooth years ago?”— the agent stops. It does not diagnose, does not pre-empt a professional judgement, and flags the thread for the dentist. It carries on with what it can resolve: booking the assessment.
Case context
Incoming message at 21:40
Active campaign with documents: fees and schedules
24-hour window opens on reply
Striker agent
WhatsApp Business
Next step
Evaluation booked; the clinical question is flagged for the doctor.
Controls it respects
- Answer only with citable facts from the documents
- Clinical questions always go to a professional
- No diagnoses, no outcome promises
The controls
In healthcare the boundary is not a matter of tone: it defines what may be said at all. That is what these three controls are for.
- Only citable facts from the documents. Price, hours and coverage come from the approved list. What is not in a document does not get answered.
- Clinical questions go to a professional. The agent recognises when a question stopped being administrative and flags the thread, rather than answering anyway.
- No diagnoses, no promised outcomes. No reply predicts how a treatment will go. The dentist says that, at the assessment.
The conversation
The full thread, with the note of what Striker did between one turn and the next:
Striker agent
Hello! Per our current fee schedule, the initial implant evaluation is free and treatment is quoted after that evaluation. We're open Saturdays 9:00-13:00.21:41Striker agent
That evaluation is for the doctor to make. I'd rather not guess at a clinical answer. Shall I book Saturday's evaluation at 10:00?21:45Fictional examples based on real product flows.
What is left
An assessment booked for Saturday at 10:00, and the clinical question flagged for the dentist — who sees it with the whole thread, not as a loose message.
The clinic did not gain a night shift. It gained the 21:40 question reaching Saturday instead of being lost.