What I check before signing off on a Planmeca deployment
If you're specifying Planmeca equipment in 2024 — a ProX intraoral unit, a ProOne or ProMax CBCT, or the Romexis imaging stack with AI-assisted features — three things get checked before anything goes to procurement. The service manual revision has to match the firmware on the unit, not just the product family. The AI imaging feature needs a regulatory clearance that covers the specific clinical claim being made. And the delivery window for original parts needs to be confirmed in writing, because 'expedited' means something different to a distributor than it does to a clinic with a Monday morning schedule.
Miss any of those three and the hardware still works. It just stops being auditable. You can't defend it in a regulatory review, you can't predict what happens when it goes down, and you can't explain to a buyer why the install passed inspection on paper but failed on the floor.
Why the service manual revision is the first thing I ask for
People assume Planmeca's service documentation is hard to find because it's proprietary. That's backwards. The reason the ProX service manual is hard to pin down is that there are multiple revisions, each tied to a firmware generation, and the distribution channel has no incentive to keep them aligned. A distributor ships what's on the shelf. If the manual on the shelf is tied to firmware 3.x and your unit is running 4.x, the torque values, the calibration sequence, and the exposure parameter tables can all be different.
That's a causation reversal most buyers don't catch. It looks like a vendor secrecy problem. It's actually a version-control problem — and version control is the buyer's job to enforce, not the manufacturer's job to solve.
In our Q1 2024 audit we flagged eleven installs where the manual revision on file didn't match the unit. Two of those had a documented torque spec off by roughly 15% against the current revision. Nothing failed. But 'nothing failed yet' is not the same as 'compliant.' I've been doing this for four years — reviewing 200-plus items a year, give or take — and the number of deployments I've had to send back for documentation alone is still the largest category of rejection. Around 30% of first deliveries in 2023, if I'm remembering the log correctly. Might be closer to a quarter. Either way, it's the top line on the rejection sheet, and it has nothing to do with the hardware.
So ask for the manual revision letter before the unit ships. It's usually on the cover page, not the title page — which is a detail that has cost me more phone calls than I care to admit.
AI dental imaging in 2024: a clearance is a claim, not a feature
The 'planmeca ai dental 2024' searches are mostly people trying to figure out what's actually cleared versus what's marketing. Fair question. AI-assisted imaging features get cleared for specific tasks — cephalometric landmarking, caries detection on bitewings, dose optimization — and the clearance covers that claim. Not the product line. Not the software platform. The claim.
This is where buyers get burned. The same 510(k) logic applies whether you're buying a dental CBCT, a blood analyzer, or a prosthetic limb component: the clearance attaches to a defined indication, and using the feature outside that indication shifts the liability onto the operator. CE MDR is stricter still on post-market surveillance for AI-enabled software, which matters if you're deploying across EU sites.
So when the rep says 'our AI reads the scan,' the follow-up question is: cleared for what indication, in which market, as of which date? If the answer takes more than thirty seconds to produce, it's a demo feature, not a validated one. Same rule applies to something like a pulse oximeter — the number on the screen is only as good as the motion and perfusion conditions it was validated under. The device doesn't become more trustworthy because the interface looks cleaner.
What 'rush' actually means for original parts
Here's the part that catches clinics. Expedited parts from a manufacturer usually don't mean next-day. They mean 'a committed date' — which is a different product. Standard is 'we'll get to it.' Expedited is 'we will confirm a date and hit it.' The premium buys determinism, not speed, and those two things get conflated in every sales conversation I've been in.
The two-hour decision I got wrong
Early 2024. A ProX went down at a site with fourteen booked patients on Monday. I had two hours to decide: original parts on a committed five-day window at a premium, or a local third-party part available in 24 hours. Normally I'd run both quotes past the biomedical team. There was no time. I went third-party on the basis that 24 hours beats five days.
The part arrived in 24 hours. It didn't fit. Two days lost waiting for the return authorization, then five days for the original part anyway. Seven days total, one rescheduled week, and a very patient practice manager.
Even after I approved the third-party order, I kept second-guessing myself. What if the fit was off by a fraction? What if the calibration drifted? I didn't relax until the tech signed off on the install. In hindsight, I should have pushed the timeline and gone original from the start. But with the schedule in front of me, I made the call with the information I had.
The assumption is that third-party is cheaper because it's the same part without the markup. The reality is that third-party is cheaper because it hasn't been through the same version-control pipeline — and for a ProX with a specific firmware generation, 'same part' is doing a lot of unearned work in that sentence.
When the cheaper route is the right one
None of this is an argument for always paying the premium. There are real cases where third-party and expedited options make sense: non-critical accessories, sites with a functioning backup unit, installations where the equipment isn't on a continuous-care schedule, and any deployment already covered by an uptime SLA in the service contract. If the contract already guarantees a response window, don't double-pay for a rush you've already bought.
And there's a real limitation on the AI side too. If the clinic doesn't have a clear clinical workflow for the AI output — who reviews it, how it's logged, whether it feeds into the treatment plan — then the clearance question is premature. You're buying capability you haven't built the process to use.
The rule I've landed on, after enough of these calls to have a default: pay for the date, not the speed. A committed date on original documentation is worth more than a fast delivery of an uncertain part, and in a review, it's the only thing that survives the paper trail.
