A Robot Just Did a Crown Prep in 15 Minutes

In July 2024, a company called Perceptive completed the world’s first fully autonomous dental procedure on a human patient. It was a crown prep, done entirely by a robot, in about fifteen minutes. The same procedure normally takes two hour-long appointments.

My first thought was: good Lord, I’m going to be out of a job. My second thought was more useful, and I’ll get to that in a moment.

How it works

The system uses a handheld 3D scanner based on optical coherence tomography to map the mouth, not just the surface but beneath the gumline and under the enamel. The AI plans the preparation from that 3D model and a robotic arm executes it.

Perceptive claims over 90% caries detection accuracy with this imaging, versus roughly 40% with conventional radiographs, and without ionising radiation. Both the World Economic Forum and New Atlas covered it as a significant milestone in medical robotics.

What’s genuinely promising

Fifteen minutes instead of two hours. That changes practice economics and patient experience in one go.
Consistent prep quality. Crown prep quality varies a lot between clinicians, and a system that produces a reliable prep every time raises the standard across the board.
Radiation-free 3D imaging. If OCT proves reliable at scale, that’s a meaningful step forward for diagnostics.

What I actually think

Someone still has to consent the patient, take the history, decide whether a crown is the right treatment, manage complications and take responsibility for the outcome. That someone is a dentist.

A robot can prep a tooth. It can’t decide whether the tooth should be prepped.

Dentistry requires a tremendous amount of responsibility. When something goes wrong mid-procedure, whether that’s an unexpected pulp exposure or a distressed patient, someone has to make a judgement call in the moment. You can’t programme that into a robotic arm.

I see this as AI being an adjunct to a skilled worker, not a replacement for one. The dentists who should pay attention are the ones who define themselves purely by the mechanical work. If all you bring is the ability to prep and restore, a machine may eventually do that part faster. But diagnosis, treatment planning, communication and long-term care: that’s the job. And it’s not going anywhere.

The reality check

This was one procedure, on one patient, in Colombia. There’s no peer-reviewed clinical data yet and no FDA clearance, which is estimated to be about five years away. The company has $30 million in funding, including from Edward Zuckerberg (Mark’s father, who is a dentist) and PDS Health, one of the largest US dental chains. That’s serious backing, but backing isn’t the same as clinical validation.

I’ve seen enough promising dental tech quietly disappear to know that a proof-of-concept is just the start. The distance between “it worked once” and “it works reliably in a busy practice on a Tuesday afternoon” is vast.

Worth watching. Not worth panicking over.

References

Perceptive. (2024). “Perceptive Completes World’s First Fully Automated Dental Procedure on a Human.” BusinessWire
World Economic Forum. (2024). “First fully automated dental surgery completed by robot.” WEF Stories
Blain, L. (2024). “Fully-automatic robot dentist performs world’s first human procedure.” New Atlas