Eleven weeks after our Planmeca ProMax 3D went into service, the finance director circled the imaging line in red. He didn't add a note. He didn't need to. Cost per scan was running well above the number in the capital request, and no explanation I gave was going to survive contact with the spreadsheet.
I'm the person who signs equipment orders for our group. I've managed the equipment budget—roughly $340,000 a year—for the past six years, and I've documented every order in our cost tracking system. I am not a radiologist and I don't pretend to be one. My job is noticing when the price of owning a tool quietly drifts away from the price of buying it.
The equipment that crosses my desk is not neatly categorized. In the same quarter I approved the Planmeca purchase, I priced a slit lamp for an ophthalmic partner and signed off on an ostomy bag order for the home-care side. Those products have almost nothing in common, except the most important procurement lesson I know: the cheapest option on paper is rarely the cheapest option in practice. The Planmeca ProMax 3D is the reason I finally stopped repeating that mistake.
The invoice was never the expensive part
We paid around $185,000 for the ProMax 3D and the AI-related imaging package. I compared it against two other systems from well-known manufacturers and asked the same questions—dose, image quality, service response times, upgrade path. Planmeca wasn't the lowest quote. It won because the clinical team thought it gave better images at a defensible dose, and because the software workflow would fit our referral patterns. All fair.
But the ROI model had one weakness. I had built it the way procurement people have built imaging models for years: machine price, finance cost, service contract, expected scan volume. The model said per-scan cost would fall quickly after the first year. (Should mention: that was the model's assumption, not the vendor's promise.) The debt service and maintenance were predictable and boring, just like they should be. The expensive part was everything that happened after the machine was switched on. It wasn't in the quote.
Problem one: nobody took the installation manual seriously
When I picture our kickoff meeting, I remember being shown the Planmeca ProMax 3D installation manual and thinking it was a technician's reference—not a document that belonged in my budget. That was the first mistake. The manual isn't just about cables and bolts. It describes the conditions that have to be true for the system to produce reliable images: room setup, power and network requirements, software configuration, and—this is the part we missed—the acceptance testing that should happen before the first patient is scheduled.
What did we miss in practice? We signed the handover certificate. The dealer's team completed the physical installation and ran through a demo. Nobody on our side verified that the items in the manual had been completed. We assumed the dealer would tell us if something wasn't right. The dealer assumed our signature meant we had checked. When a service technician came out months later about an unrelated complaint, he asked where the baseline acceptance results were. There was a long pause.
This unit wasn't accepted until today, was it?
He wasn't rude. He was just stating the facts. We had installed a $185,000 scanner and skipped one of the most important quality steps, not because anyone was negligent but because both sides used the word accepted and meant different things. I said we expected everything in the manual to be completed before go-live. They heard that the usual post-installation signature would be fine. The result was a scheduling gap, a service visit, extra paperwork, and no reliable baseline for image quality until weeks later.
Let me rephrase that so it lands properly: the installation manual is a contract document, not a brochure. You should treat it as a checklist that the vendor has to prove they've completed before you sign anything.
Problem two: Planmeca AI dental imaging didn't fail. We failed to give it a workflow.
Once the acceptance issue was sorted, the scanner worked well. The next problem showed up in the per-scan numbers and it was quieter. The AI package—let's call it what the marketing materials call it: Planmeca AI dental imaging—was not saving us the time the business case expected. I started asking why.
The answer wasn't technical. The feature was functioning. The problem was that two of the four clinicians didn't trust it, so they reviewed every scan the way they had before, and then checked what the AI had flagged. That is not using AI to save time. That is doing the old job and adding a new task on top. One of the dentists put it to me directly: if the software flags something, I still have to look at the image and decide whether the flag is right. Why would that save me time?
Fair question. It saves time only when the workflow is redesigned around it. The AI module does not replace the clinician; the interpretation still sits with the dentist. The time saving is supposed to come from the fact that the software has already done some of the routine work. But if every flag is re-checked from first principles, there is no saving. The technology was an evolution, not a revolution, and I had presented it to the team like a revolution. (I should add: the clinicians were right to be cautious at the start. The rollout was my process failure, not their stubbornness.)
The phrase Planmeca AI dental imaging sounds more autonomous in a brochure than the daily workflow really is. Per FTC guidance on advertising, claims about what a product does need to be truthful and substantiated (see ftc.gov/business-guidance/advertising-marketing). It was a useful frame for a buyer: ask the seller what exactly the software does, because the brochure will not tell you.
It was also a reminder that imaging purchases come with an education component. Around the same time, one of the newer associates asked a simpler question: what is fluoroscopy, and could the ProMax do it for certain joint cases? The short answer is that fluoroscopy is real-time x-ray, while a CBCT like the ProMax 3D captures a 3D volume in a single rotation; they answer different clinical questions. What mattered was not the physics question but the fact that the team was still mapping what the new system could and couldn't do. If you skip that mapping, you end up with a very expensive panoramic machine.
The cost of quiet underuse
When I finally put the year one numbers together, not one line item pointed to hardware failure. The extra cost came from three places: the service visit and delays caused by our missed acceptance testing, training that happened after go-live instead of before, and the months where half the clinical team was doing double reading work because they didn't trust the new workflow.
I don't have hard data on what that distrust cost in clinical terms. I wish I had tracked it more carefully. What I can say anecdotally is that the gap between the doctors who used the AI tools and the doctors who didn't was visible in the schedule: same number of scans, different time per scan, no difference in patient care. The finance director didn't need to know how to read a CBCT. He just needed to see that we had bought a time-saving tool and then gave the team no permission or process to save time (note to self: measure the before and after, not just the after).
This part matters because it changes the calculation. If you buy a Planmeca ProMax 3D and use it as a slightly better panoramic unit, you'll still get good images. You just won't get the financial outcome that justified the purchase. The machine doesn't force the workflow to change. The team has to change the workflow, and that is a project, not an installation event.
What I would do differently
If a second clinic asked me to handle the same purchase tomorrow, the equipment decision would stay the same. The process around it would look completely different.
- Read the Planmeca ProMax 3D installation manual before signing the purchase agreement, not after the service visit.
- Make the final payment contingent on documented acceptance tests, with the manual's checklist attached to the purchase order.
- Budget for training and workflow redesign as if they were hardware line items. They are. (circa 2025, I would still put this at 5 to 8 percent of the total project cost.)
- Set a clear protocol for how AI findings are handled before the first scan, then review the protocol after 90 days with the actual users.
The fundamentals haven't changed: a dentist still interprets every image and is responsible for the outcome. What has changed is the expectation that software will quietly make the practice more efficient without anyone updating the way the practice works. That assumption is outdated. I found out the expensive way.
We kept the scanner. We changed our process. By year two, per-scan cost finally looked like the model I originally built. The red circle disappeared. The machine was never the problem.
