The 11pm Patient Israeli Clinics Keep Losing

At 11 o’clock at night, somewhere in Israel, a woman with a toothache picks up her phone. She isn’t calling anyone — nothing is open. She’s scrolling through Google Maps, tapping on dental clinics, looking for one where she can do something right now: book a slot, ask a price, anything. The first clinic that lets her act at 11pm gets her as a patient. The other nine get nothing, and never know she existed.

I build automation systems for small Israeli businesses, and clinics are where this pattern is most brutal. Not because clinic owners are behind the times — most of them run tight, modern practices — but because the appointment book is the entire business, and the appointment book is managed by humans who go home at six.

Two numbers explain almost everything about the economics of a clinic. The first is how many people try to reach you when nobody can answer. The second is how many people book an appointment and then simply don’t show up.

The 23 percent problem

Start with the second number, because it has decades of research behind it. A systematic review by Dantas and colleagues, published in Health Policy, synthesized more than a hundred studies of medical appointment attendance going back to 1980. Across specialties and countries, the average no-show rate lands around 23 percent. Nearly a quarter of booked appointments end with an empty chair.

The same review found something more useful than the headline number: the strongest predictors of a no-show are long lead time — the gap between booking and the appointment itself — and a history of previous no-shows. In other words, the patient who booked three weeks ago and heard nothing since is exactly the patient who won’t come. Not because they decided against it. Because life happened, and the appointment quietly fell out of their head.

Israeli clinics I’ve sat with recognize this instantly. One clinic we worked with — I’ll keep it anonymous, as with all our client stories — was running at roughly 30 percent no-shows before we touched anything. On a day of 20-something booked treatments, six or seven chairs sat empty. The front desk knew there was a waiting list of people who would have killed for those slots. There was simply no mechanism fast enough to connect an empty 14:30 with a person who wanted a 14:30.

What the front desk actually does all day

Before that project, we shadowed the clinic’s reception work. The secretary was answering the same five questions — where are you located, what are your hours, how much does a checkup cost, do you take my insurance fund, when’s the next opening — around 50 times a day. Every one of those calls interrupted something: intake, billing, an actual patient standing at the desk.

And then there was the 11pm woman. The clinic’s phone log didn’t show her, because she never called. But the clinic’s WhatsApp did — messages arriving at eleven, at midnight, sometimes later, answered the next morning at nine. By which time, often enough, the toothache had found another clinic.

This is the part that surprises owners most when they finally measure it: the leak isn’t dramatic. It’s a slow, silent drip of people who tried you first and got silence for ten hours.

What changed

In March 2026, we deployed a WhatsApp bot for that clinic — the kind of project boutique automation studios such as Achiya Automation build for small businesses across Israel. The architecture is almost boring: WhatsApp on the front, an automation engine in the middle, Google Calendar at the back. The bot answers within about a minute, at any hour. It shows real open slots, books directly into the practitioners’ calendars, and — this is the part that moves the needle — runs the reminder cycle on its own.

The reminder cycle is where the no-show number lives. A message goes out the day before: your appointment is tomorrow at 10:00, confirm or reschedule with one tap. Across our deployments we’ve added a second reminder two hours before the appointment, and that second touch consistently cuts last-minute no-shows further — the day-before message jogs the memory, but the two-hour message catches the conflicts that only appeared that morning. When a patient does cancel, the bot immediately offers the freed slot to the waiting list, one person at a time, until someone takes it.

At that clinic, the no-show rate dropped roughly fivefold. Not to zero — humans remain humans — but from nearly a third of the book to a level where an empty chair is an event rather than a Tuesday. The secretary still works full days. She just spends them on patients instead of repeating the clinic’s address into a phone.

Notice what this did to the research finding about lead time. The bot didn’t shorten the calendar — people still book weeks ahead for some treatments. What it shortened was the silence. A patient who gets a confirmation at booking, a reminder the day before, and a nudge two hours out is never three quiet weeks away from their appointment. The lead time is the same; the abandonment window is gone.

The math nobody runs

Clinic owners tend to evaluate this kind of system as a technology purchase, which is the wrong frame. The right frame is: what does an empty chair cost? Take a modest private practice — eight to ten treatments a day. At a 23-to-30 percent no-show rate, that’s roughly two empty slots daily. Price them at whatever your average treatment is worth and multiply by a working month. For most Israeli clinics that’s a four-figure monthly loss in shekels, often five — quietly, every month, for years. A system that claws back most of it pays for itself in weeks, not quarters.

And the 11pm patient is pure upside on top. She was never in the book, so she never showed up as a loss in any report. The clinics that answer her at 23:00 are not out-marketing their competitors. They’re just awake.

None of this requires a hospital IT budget or a development team. That’s the actual news here — not that the technology exists, but that it has quietly become small-business infrastructure, the way a website was in 2010 and a credit-card terminal was in 1995. The clinics adopting it aren’t early adopters anymore. They’re just the ones who measured.

So here’s my question for clinic owners and managers reading this: of tomorrow’s appointment book, how many patients do you actually expect to walk through the door — and when was the last time you counted the ones who didn’t?

Achiya Cohen builds WhatsApp and workflow automation for small businesses in Israel. He writes about the practical economics of automation for SMBs.