What Is the Future of AI in Dental Laboratories?
Chuck Stapleton
What's the future of AI in dental laboratories? That's the prompt I posed to ChatGPT, one of the most popular chat AI platforms on the internet. I wanted to compare notes, essentially, and it actually brought up something I had not considered—more on that later.
Artificial intelligence (AI) has made its way into almost every conversation in our lives. Whether in business or social conversations, there is a good chance someone has mentioned AI to you. But what does that mean for dental laboratories and your business? What about dentistry as a whole?
Here is a quick story to illustrate where I believe things could be headed.
The AI Crown
The story begins with Patient Smith coming in for a regular cleaning, and Dr. Miller's assistant taking a digital x-ray. The assistant feeds the x-ray images into an AI diagnostic system the office uses to detect caries and other clinical issues. This particular system is paid for by a dental laboratory partner. Why would any dental laboratory do that? The AI helps the dentist catch any decay they may not have noticed, which can then lead to more work for the laboratory.
The diagnostic comes back and informs the dentist there is decay on one of Patient Smith's molars—something that the dentist and assistants did not catch when they evaluated the images the first time.
A few days later, Patient Smith returns to the office to have a digital scan performed. Dr. Miller uses an intraoral scanner to capture the patient's dentition and sends it off to the laboratory. The scans are checked instantly by the laboratory's AI software, and some news is delivered that could have been devastating under other circumstances: The dentist missed scanning the mesial contact of the tooth adjacent to the molar that is getting the crown. Fortunately for Dr. Miller, the patient is still in the chair and the missing area can be rescanned, without the drama of needing to schedule a reappointment. After that is complete, some undecipherable case notes are added to the file and it is sent to the laboratory. The laboratory's AI system is able to figure out what Dr. Miller wrote, and it also suggests a better treatment plan than what the dentist had recommended. The office is contacted via text message and a different product is suggested. Dr. Miller agrees and signs off to go ahead with the change. A new delivery date is provided at the end of the conversation.
By the way, this all happens via text, within an hour of the case being sent to the laboratory. Amazingly, no one from the laboratory needed to monitor or participate in the conversation.
The case enters into production, and an AI design is recommended based on the particulars of the patient's dentition, but also with considerations for Dr. Miller's case history, education, training, geographic location, and age. This design will provide the highest chance of case acceptance for the dentist and the patient. The case is then scheduled based on the available technicians and equipment. However, one particular mill has been showing signs of potential failure, so the AI system removes that mill from the production process and informs the laboratory manager it needs some maintenance. The system then replans the production around that mill being down for repair.
After the design is approved by a technician, the CAM software takes over, and that is also driven by AI. The crown is nested into a disc and the model is nested on a 3D printer. Both situations result in maximum use of space and materials. This has allowed the laboratory to save a significant amount of money on material usage and equipment purchases.
Next, however, tragedy strikes! While the crown is being moved from the milling machine to a sintering oven, it is dropped, and it breaks. Screams are heard and tears are shed, but eventually, everyone composes themselves and does what they were trained to do in this situation: Enter a case failure into the AI system. It then re-balances production and moves the case ahead of others to meet the dentist's due date that was set.
This is a good time to note the importance of what had been done months prior to today to ensure that a case like Patient Smith's could be handled as efficiently and effectively as possible. The laboratory manager had hired a consulting company to bring in its AI system to determine the most optimal layout for the facility. A recommendation was made based on equipment, square footage, products available, technician skill levels, and technician pay. This is what allows them the flexibility to reroute a problem case through production when necessary.
In addition to the efficient layout, all the necessary materials are available and in stock, thanks to the AI inventory management system that predicted what would be needed this month based on seasonality, sales input, patient demand, information from the ADA, and other public resources. It even went as far as ordering extra glaze from a particular supplier due to the owner of the company being caught in scandal that sent shockwaves through the dental world. The laboratory was advised to stock up on 6 months' worth of stain and glaze from this supplier, which turned out to be a life saver because the glaze supplier ending up ceasing operations.
Back to Patient Smith's case: While the case is in production, Dr. Miller's office manager sends an email to the laboratory saying that the patient wants to come in a day earlier. "Can it be delivered any sooner? The patient has a wedding they can't miss, that is also the same day as their anniversary, children's birthdays, grandchild's college graduation, and headshot photos, all in the same day!" The customer service system, powered by AI, responds to the email almost instantly (while maybe rolling its eyes), stating that the case will be delivered one day ahead of schedule according to where it is in the process. It even checks for potential weather impacts that could slow the case down when it leaves the laboratory, but fortunately does not identify any problems currently.
As the case reaches QC, a technician scans the final prosthetic with a 3D scanner and uses a color matching device to obtain the shade. This is fed into the QC system, which, if you have not figured out by now, is also AI-powered. It checks to see if the restoration fits within the tolerances of what this particular dentist would accept. The system gives the go-ahead to deliver the case. It is packed up and sent to the dental office.
Tragedy strikes again, though! After delivery, the office manager notices a billing error on the case. The office was accidentally assigned to the wrong price list and was overcharged. Discussions ensue about calling the police, the FBI, and hiring lawyers, but cooler heads prevail about the $10 overcharge. Dr. Miller calls to complain about the issue and the laboratory manager immediately apologizes, gives the account a credit, and logs a complaint in the system. Jail time and lawsuits avoided.
The following day, Dr. Miller's account is flagged by the AI system as being at a high risk of leaving the laboratory. Based on other case issues the dentist has had and the most recent billing error, a sales person is being sent to their office to save the account before they look elsewhere. The sales person was not told any of this by a manager because they were informed automatically by—you guessed it—AI.
Thus ends the story of the AI crown.
Discussion
Some elements of this story reflected applications of AI that I know are being utilized already in dental laboratories; some are merely being talked about at this point, and a few are original ideas. Oh, and that suggestion ChatGPT made that I had not thought of when I asked it about the future? Predictive maintenance.
However, all of this relies on AI systems having access to enough data—accurate data. This may be a bigger hurdle than the technology itself. Perhaps dental laboratories should look at combining all of their data together to create a better outcome for all? That would be a massive undertaking to convince the industry to work together in this way.
What will be the cost? AI development and systems require significant investment. Those that make that investment would be expecting to get a return in licensing fees or by maximizing the use of those systems to create a competitive advantage.
I talked with Ben Schrauwen, Senior Vice President and General Manager of Oqton, whose software is AI-powered, and he nicely summarized what AI will bring to us: It is the commoditization of expertise. I do not think I have heard anyone summarize AI's impact on the world so succinctly.
How far are we away from this Utopia—or nightmare, depending on your point of view? My guess would be no more than 10 years.
And it would not be an article about AI if I did not include some AI-generated content. I generated a few images using Midjourney, an AI platform for creating art and photographs. Some very basic prompts were used, which are in the captions of the images (Figure 1 through Figure 4), and no time was spent refining the AI. You will notice a few things that are not quite right, and these were left in intentionally to show some current limitations. However, if more time was spent with the AI engine, they would be nearly indistinguishable from real photographs.
Lastly, the following is what ChatGPT produced when I asked it to write a poem about dental laboratories and AI:
In dental labs, where wonders thrive,
AI and artistry doth collide.
With precision and digital grace,
Smiles are crafted in a seamless embrace.
Algorithms whisper, scans they explore,
Mapping teeth like never before.
Guiding hands with knowledge profound,
Creating dentures, beautifully crowned.
AI's wisdom, a trusted friend,
Enhancing skills, a helping hand.
Together they weave a masterpiece,
Transforming lives with smiles increased.
In dental labs, a future bright,
Where AI and art take flight.
With technology's aid, they stand tall,
Creating dental wonders for one and all.
About the Author
Chuck Stapleton
Vice President and General Manager for Dental
3D Systems Corporation
