Where AI actually lands in the dental industry — a map from a broker who's been paying attention.
I've spent the last while trying to answer one question honestly: where does AI actually take our industry? Not the hype — the mechanics. Who does what today, and what can a machine now do instead. I automated my own brokerage this year partly to find out firsthand, and this memo is what I found.
I'm not selling anything here, and I'm not warning you about a product I happen to have. AI isn't coming because of me or anyone like me. It's coming on its own. I just think this is the conversation our industry should be having, and almost nobody is having it out loud, so I wrote it down — with the numbers, so you don't have to take my word for anything.
The short version before the long one: the people who engage this technology are going to be fine, and the people who wait for permission are going to get passed. That's not a threat. It's a pattern, and the pattern is old.
On the human touch — because it's the first thing everyone reaches for.
I get it, and I don't say any of this lightly. Real people get displaced in these shifts, and the anxiety is legitimate — because it has happened, over and over, to real people. The telephone switchboard operator was once one of the most common jobs in America for young women; the job effectively no longer exists. Elevator operators — gone so completely we forget it was ever a job. Travel agents were somebody's mother, somebody's friend; the internet cut the profession roughly in half and turned the survivors into specialists. Bank tellers watched the ATM and then the phone absorb most of what a branch did. The video store, the classified-ad desk, the retail rep that e-commerce quietly made unnecessary. And the bookkeepers — the spreadsheet was the first machine that could reason over numbers, and an entire profession reorganized around it within a decade.
Notice two things about that list. First: every one of those was a relationship business, and everyone in them was sure the human touch would hold. Second: the world didn't end — it reorganized, and the people who moved early came out ahead. The plain business truth underneath all of it: cost efficiency beats the human touch every time — until the moment the human touch is genuinely needed. The whole game is knowing where that line falls. The human touch was never the paperwork, the scheduling, or the math. It's the hand on a scared patient's shoulder and the name that stands behind a decision. Everything on the wrong side of that line is already leaving. Everything on the right side just became more valuable.
"It can't be done" has already entered the chat.
This March, the largest practice-transition company in the industry published a warning to dentists about relying on AI in practice sales. Their words, verbatim: "There is no universal marketplace for dental practice sales, and there is no simple formula that guarantees an accurate valuation or a successful transition" — and AI "is not a substitute for thoughtful guidance, real-world experience, and a team that understands both the business and personal sides of this decision."
Two honest things about that. First: they're half right, and it's the half this memo agrees with — judgment, context, and accountability are exactly what survives. You'll notice they're describing the choke person. Second: study the pattern. When the incumbent takes the time to publish "it can't be automated," the clock has already started. The video store had opinions about streaming, too.
So here's the frame I'd offer instead of the fear or the denial. Stop picturing a robot coming for the profession, and start picturing an employee — the best, smartest employee you will ever hire. It has read every deal. It never misses a Monday. It's never hungover, never out sick during your closing week, never distracted by its phone. Sure, there are edge cases where you step in — that's true of every employee you've ever managed, and this one takes feedback instantly and never makes the same mistake twice. You wouldn't turn down the best hire of your career because the old guard published a blog post. The only real question is who this employee works for: you, or the firm across the street.
So here's the map, silo by silo — with receipts.
Lending. Industry analyses of automated underwriting report loan decisions that took days now clearing in minutes, at 35–50% lower processing cost per file — and the AI-lending market is projected to grow from roughly $110 billion to $2 trillion by 2037. Follow the mechanics: behind the dental banker you know sits a chain of processing, underwriting, and credit, and nearly all of it is pattern recognition — the one thing these systems now do better than we do. Underwriting, in banking and in insurance, should already be automated; where it isn't, that's a countdown, not a debate. The market itself turns fluid: software that shops every lender's real appetite in seconds. What endures is the banker at the front who can walk a hard deal into the room and put their name on it. Fewer of them, each doing the volume of ten.
Lending — what automates: underwriting · processing and file assembly · document collection · rate and program shopping · first-pass credit analysis. Citi analysts put over half of banking work in the high-automation column.
What survives: the banker who owns the outcome and can carry an exception through the room.
Accounting. Start with what the biggest firms in the world are already doing. The Big Four have poured a reported $9 billion into AI — and PwC cut about 1,500 U.S. positions in 2025, KPMG trimmed its U.S. audit workforce, and Big Four graduate hiring fell 44% year over year. Read that again: the largest accounting firms on earth are replacing their incoming class with software. Analysts now put 42% of finance and accounting tasks in the automatable column, while the AICPA reports a decade-long slide in new CPA candidates. Both ends of the pipeline are narrowing toward the same point, and that point is the whole profession in one sentence: what a client needs is one person who signs the return. The layers beneath that signature — extraction, categorization, add-back schedules, normalized statements — are minutes of machine work now. The sharpest CPAs I know aren't fighting it; they're quietly building smaller, faster practices around exactly that: one name, unlimited compute, no toll booth between a client and their own information.
Accounting — what automates: data extraction from returns and statements · transaction categorization · add-back schedules · normalized financials · first-draft workpapers · the monthly "access" subscription. Roughly 42% of the work, by current industry estimates.
What survives: the partner whose signature carries the liability — and the judgment calls underneath it.
Legal. Here's a study that should get more attention than it does: back in 2018 — before modern AI, seven years ago — LawGeex put its contract-review system against twenty experienced corporate attorneys, including lawyers from Goldman Sachs and major global firms, on five NDAs. The AI hit 94% accuracy in 26 seconds. The lawyers averaged 85% — in 92 minutes. That was the old technology. The dental-specific attorneys picked a niche early and were right — niches endure — but the work is dividing: document assembly completes in moments; judgment on the genuinely novel terms, with a name behind it, does not. The attorneys who thrive will run several times today's volume on the same headcount, treating the machine as the best associate they've ever had — reads everything, works around the clock, never has a rough Monday. The craft moves from drafting to directing.
Legal — what automates: contract assembly and markup · document review · precedent search · first drafts of routine filings. Goldman Sachs analysts have estimated roughly 44% of legal tasks are exposed to automation.
What survives: judgment on the genuinely novel terms — and the name that stands behind it at closing.
Consulting. By this year, industry surveys show a majority of dental practices adopting or planning automation in their operations. So let me make this one concrete, because your office may already own the first piece. If you have voice charting, look at what's actually installed: a system that hears. Soon a practice will summarize its own day — every operatory, every front-desk call, every case presentation — and tell you how it did. Ask the best owners which operatory is their favorite, and more than one has told me: the middle one, "because I can hear the front and the back while I work." Every great owner runs by ear. Now imagine hearing everything, every day, organized by a tireless analyst before lunch. That's what practice consulting becomes. The binder and the twelve-month engagement were the organizing of your own information — and organizing information just stopped being scarce.
Consulting — what automates: the practice analysis · overhead benchmarking · fee studies · the binder · the report-writing that filled a twelve-month engagement.
What survives: the few operators distributing real systems — and the trust to be let inside a practice.
Supply. Dental supply has been waiting for its Amazon for twenty years and mostly got lucky. The field rep is a communications layer between an office and a catalog, and the cost of that layer — salary, car, benefits — sits inside your margins, the same way it did in every retail category e-commerce has already flattened. We've watched this sentiment play out inside dentistry before: a generation of owners wouldn't bring specialty work in-house because they didn't want to take food off a friend's table. Honorable — and it cost them a fortune. Loyalty to a distribution layer is the same decision in a nicer jacket.
Supply — what automates: ordering · reordering · price comparison · product research · the catalog visit itself.
What survives: service on the equipment that's bolted to the floor — and not much else in the middle.
Brokerage — my own chair. I saved my own profession for last on purpose — I'm not willing to say any of this about anyone else's business without having done it to mine. Listing marketing, buyer matching, valuation modeling, follow-up, scheduling — I automated all of it in my own shop this year, and it works. The broker whose entire function is a listing agreement and a lockbox is finished. What's left is regional: a few accountable dealmakers who leaned into systems early and now answer in minutes, prepare buyers before the first call, and never drop a thread. The one with the systems eats the region — not because they're smarter, but because they're present in a way no unassisted person can be. And to be clear about what "leaning in" means: the AI doesn't take over a business. You still have to analyze your own operation, find where the machine cuts in, and direct it. It amplifies an operator. It doesn't replace the operating.
Brokerage — what automates: listing marketing · buyer matching · valuation modeling · follow-up · scheduling · buyer preparation. I know, because I automated each of these in my own practice this year.
What survives: the accountable dealmaker at the table when the deal is on the line.
DSOs — the uncomfortable page. The numbers first: DSO affiliation is projected to hit 39% of U.S. dental offices this year, up from 23% in 2024, and the dental-AI market itself is on track to nearly triple by 2033. My honest read, and independents won't like it: the consolidators are positioned to win this wave. The small practice never lost patient-by-patient — chairside, the independent wins that fight every day. The fight is at scale, and scale is exactly what changed: one integration deploys across four hundred chairs while a solo office is still reading the demo email, and the margin gap between AI-run operations and everyone else is what pulls smaller practices into the bigger model. Do I think dentistry should be run by groups? No. It's one of the most personal services in medicine, and I'd protect the private practice with everything I have. But I don't see a way around what's coming for anyone who sits still. One warning while I'm here: a lot of money is about to be lost to salesmen with "AI" on the deck and nothing behind it. If they can't show you their own operation running on what they sell, keep your checkbook closed.
Here's the twist, though — the reason I'm writing instead of shrugging. For the first time, the same weapons are free. The back office only a large group could afford — the analytics, the follow-up engine, the phone answered at 2 a.m. — a solo practice can run today for close to nothing. So it cuts both ways: consolidation wins by default, or the small owner picks up the big group's infrastructure and keeps the one thing a group can't manufacture — a name the patient personally trusts. That window is open right now.
What every silo has in common.
Run them back and the same person is left standing in each one. Not the smartest. Not the cheapest. The one you call at 4:50 on a Friday when the wire didn't land.
Over more than two decades I've narrowed to a handful of bankers and a handful of dental attorneys, in a country full of both. Why so few? Because they're my choke people — the point in a process where accountability stops being diffuse and becomes a name. Every business runs on customers who never appear on an org chart: the escrow officer is my customer, and I'm hers. When something breaks, a specific person owns it to me, and I own it to my client. That chain of accountability was always the real product. Most of what surrounded it was packaging.
You can't sue a chatbot. You can't look a language model in the eye at a closing. Accountability can't be generated — only staked, by a person, with their name and their next deal on the line. So no — AI doesn't erase the professions. It erases the middle of them: the processors, the gatekeepers, the couriers of information, everyone between a client's data and a client's decision. What remains, in every silo, is one accountable name on top of enormous leverage, where there used to be one name on top of forty salaries. The pyramid doesn't collapse. It inverts.
So what does a dentist actually do about it?
Your profession is safe in the way that matters. Nothing in this memo holds a handpiece or earns the trust of a frightened patient, and the ethics that built your practice still decide everything. But safe doesn't mean still. The honest short list, from where I sit: engage it this year, in small real ways — not a pilot you forget about. Run the analysis on your own practice: what are you paying a layer to do that a machine now does for free? Choose your advisors by one test — when it goes wrong, whose single name is on it? And stop waiting for a canary.
The last number, and the one that matters most: the ADA's own data shows roughly half of practices have now tested an AI tool — and fewer than one in nine has genuinely embedded one. Nearly everyone is dabbling. Almost no one has rebuilt. The distance between those two numbers is the next five years of this industry, and which side of it you're on is still, for a little while longer, a choice.
I don't have all of this figured out, and I'm not pretending to. I'm a broker who's been paying close attention, because these are the people I work with every day. This is where I think it goes. I'd rather we see it coming together than one at a time, too late.
— Jonathan Ingalls has brokered dental practice transitions for more than two decades.
Sources & further reading: Multimodal — automated underwriting time/cost data · Timvero — AI lending market projections · TheStreet — Big Four AI spend, hiring cuts · NCS — 2025 Big 4 layoffs (PwC ~1,500, KPMG audit) · Accountancy Age — graduate hiring down 44% · CFO.com — AICPA CPA-pipeline data · Artificial Lawyer — LawGeex study, 94% vs 85% · Patientdesk — practice automation adoption · Duckett Ladd — DSO affiliation 39% vs 23% · Ricci — ADA AI adoption survey (47% tested / 11% embedded)