By Dave Salciccioli, Co-Founder and CEO, Relay.
I've been to my share of dental conferences. Most run on the same formula: a big hall, a crowded expo floor, and a string of keynotes about where the industry is headed.
The Overjet AI Summit in Scottsdale was a different kind of event. It was small and intimate, and the room was full of DSO operators, the people who actually have to make this stuff work across dozens of locations. Chris Thompson, Steve Jensen and the Overjet team deserve real credit for the agenda. They built it around doing instead of watching. Day one had a hands-on lab on voice AI. Day two sat people down with their own data, or demo data, to build something they could take back to their offices. At the opening reception, a board asked one question: if AI could solve anything for your practice, what would it be?
That set the tone for two days. The sessions were practical, the hallway conversations were valuable, and nobody was asking what AI is anymore. They were asking how to make it work in their practices to increase same store growth.
It was clear at this event that dental AI doesn't have a product problem anymore. Vendors can build almost anything, and they're shipping it faster than any of us can keep up with. What's scarce now is the ability to get it working inside a practice.
The bottleneck has moved
For most of dental tech's history, the constraint was how fast a product could add new features. Features took years. Integrations took longer. If you wanted something, you waited on a roadmap.
Wardah Inam, Overjet's founder and CEO, opened her keynote with a story that showed how much that's changed. The Friday before, driving back from dropping her son at school, she had an idea for a product launch. She wrote the script in the fifteen minutes before her first meeting and had AI build the video while she sat in it. "By the time I was done with the meeting, I had the first draft."
A conversation with Katie Poulson from Henry Schein One pointed the same direction. By her account, the team has shipped more in the last nine months than it did across the previous nine years.
Nine years of output in nine months. When building gets that fast, the pressure moves downstream, to everything that has to happen after a feature ships.
That's where the bottleneck sits now: in how quickly a DSO can take in new tools and drive real adoption across its practices. A vendor can ship a feature in a week. Getting every office to actually use it runs on a much slower clock, and most groups weren't built to absorb change at this speed.
The adoption gap is getting wider
You can see that clock from the operator's side. Rezwan, CEO of Imagen Dental Partners, named the limit on a panel later that morning: "A practice can only absorb so much innovation."
Technology has always moved faster than adoption, he said, but that gap has grown exponentially. Every new feature competes for the same things inside a practice: a doctor's attention between patients, a team huddle that's already full, an office manager running the schedule and the P&L at the same time. Push too much through at once and nothing sticks.
So his team has gotten a lot more discerning about what they take on: fewer tools, chosen on purpose and rolled out well. When vendors can ship at the speed Katie described, saying no becomes one of the most important things a DSO leader does.
This isn't only a vendor problem or a DSO problem. Vendors are built and funded to ship and activate. DSOs are built to run practices. Nobody owns the work in between: the workflow changes, the training that sticks, the check-in at day 90.
AI theater, and the honest alternative
Rezwan also had a name for what owners are pushing back on right now from vendors. For the last eight months his team has been asking one question: "What is AI theater versus what's actually creating value?"
Right now, a lot of vendor energy goes into what happens in front of the curtain. The demo lands, the room claps, and later you find out much of what you saw is still on the roadmap. Owners want to see behind the curtain: what works today, in how many offices, and with what results.
I don't want to hear another vendor talk about what is in their roadmap. I want to know what is here today.
That was Paul Smith from SGA, saying it about as plainly as you can.
My bet is that over the next few years, the vendors who are straight with customers win. Here's what works today, and here's what it will take to get it working in your offices. That kind of honesty is rare enough right now that it'll stand out on its own.
Start with the operation
Vince Dilley, CEO of Platinum Dental Services, was the clearest voice all day on what needs to happen first to be able to leverage AI. He didn't open his session with AI. He opened with how he measures a great DSO: at least 10% same-store growth year over year, with corporate overhead under 8%.
Only then did he get to AI. Platinum is connecting all of its data sources and building agents on top of them, and he said the hard part shows up fast. "The challenge very quickly becomes prioritization." His team's name for AI theater is "the cool factor." He was clear though that every idea has to move one of three things: financial performance, clinical outcomes for patients, or enhancing the support Platinum gives its providers. If it doesn't, "then it's cool, but it's not something we prioritize."
That's the order that works. Build the operation, then let AI carry it further.
What's next for DSOs
Over the next 12 months, I expect a lot of our tools to start converging. Each platform keeps expanding sideways to cover more of the practice, and most groups will end up with access to roughly the same AI capabilities. Rezwan expects the same of the data itself: "I believe that the scoreboard is going to be more or less commoditized." The tools stop being the differentiator. What a group does with them becomes the whole game.
Rezwan put the opportunity plainly: "The reality is that all organic growth that we do is inside our four walls already." The groups that capture it will be the ones with the discipline to get four things right.
Operations first. AI makes a well-run practice better and a messy one messier. Know your workflows, your data and your numbers before you layer anything on top.
Discipline about what to build and integrate. Every group can now build agents on its own data, and every vendor has a new feature to connect. Hold each one to Vince's test: does it move financial performance, clinical outcomes or the support you give your providers? If not, it waits.
Smart decisions about what to buy. Put a business case in front of every purchase, with value, effort and ROI decided up front. Imagen has 20 to 25 leaders working through use cases this way before anything gets piloted.
A real plan to drive adoption. This is the one most groups skip, and it matters most. Someone has to own the adoption, separate from whoever owns the contract. Measure adopted, not activated, because logins aren't outcomes. And take a real look at day 90, when the support tickets go quiet and you find out whether people changed how they work or quietly went back to the old way.
I've said before that a fancier car doesn't make you a better driver. The industry just got handed a lot of fancier cars at once. The groups that pull ahead will be the ones willing to do the unglamorous work of teaching their teams to drive them.
If your group is sorting out which of these tools are worth the effort, or how to get the ones you already own adopted, I'd be glad to talk it through. Connect with me on LinkedIn, or start a conversation with the Relay team.
Want this analysis for your organization? Explore Industry Intelligence →