I spent this week on the West Coast with DentAI. Los Angeles first, then San Francisco. Talking to dentists, to founders, to people building companies across the dental industry.

And I have to be honest with you about something I saw everywhere: a lot of people in this space push AI at you for one reason. They have a product to sell you. Every problem magically becomes an AI problem the moment someone has an AI solution priced for it.

Our role here is the opposite. Highlight what's genuinely worth it, and just as importantly, tell you clearly: at certain moments, AI is useless to you.

This edition is about the most expensive confusion I saw all week. And it starts with a story.

Two practices, same street. Both bought "AI" this year.

Come back eighteen months later. One is hiring a second associate. The other is staring at holes in the schedule.

When a practice adopts AI, it almost always starts with admin pain. And I get it, it's the pain you feel every single day. Notes, claims, phone tag. That AI works, by the way. It really does cut the workload, often by 20-25%.

But think about what you're actually buying.

Say AI frees up 6 hours a week of your assistant's time. Great. Are you going to fire them? Of course not. You need them, and you'll need humans in your practice for a long time to come. So the payroll doesn't move. What you bought is capacity, not savings, and the value of that capacity caps out fast: a few thousand dollars a year of labor value that you keep paying for anyway.

The ceiling is built into the problem itself. The most admin AI can ever give you back is the admin time you were losing. Not one dollar more.

Revenue isn't infinite either, a practice is a practice. But the room above your head is a completely different order of magnitude, especially on profitability per hour of chair time, the number almost no dentist actually calculates.

When nobody answers, only 14% of new patients leave a voicemail. The other 86% call the next practice on their list.

Read that again: you paid Google to make your phone ring, and you delivered the patient to your competitor.

Should it be done with AI? Well not always. It’s not necessarely the right move.

Ask any dentist their acceptance rate, the answer is "around 90%." The tracked data says many practices run up to 30 points below their own estimate.

Do the math on your own practice: going from closing 4 cases in 10 to 6 in 10, on a $1M practice, is about $160K a year. No new patients or ad spend. Just patients you already convinced to sit in your chair.

Average dentist production: $475-575 an hour. Top performers: $700+. Same chair, same hands.

The difference is not speed. It's which treatments happen, in what order, at what profitability. And no note-taking tool on earth touches that number.

Revenue solves the admin problem. Admin never solves the revenue problem.

One removes friction from the practice you have. The other builds the practice you don't have yet.

AI is not the point.

I heard it again this week, from dentists in LA and from founders in SF who were honest enough to admit it: sometimes the best ROI on your phone line is not a voice agent, it's a good human who picks up on the second ring. A patient calling with a toothache at 2pm doesn't want to negotiate with a bot. They want to feel that someone cares that they're in pain. For a lot of practices, that person converts better than any AI, and the patient walks in already trusting you.

So the question is never "can AI do this?" It's "what does this moment need: availability, or empathy?"

And the uncomfortable truth:

If your treatment presentations confuse patients, automating your follow-up just confuses them faster. If your schedule is chaos, an AI booking layer books chaos, very efficiently. AI is a lever. It amplifies whatever management is already there, good or bad.

What it can do, used deliberately, is help you build the foundations: see your real numbers instead of your estimated ones, catch the leaks you didn't know existed. Foundations first. Amplification second.

Your patient market is finite. Every new patient in your area ends up somewhere.

DSOs, projected to hit 39% of US practices this year (up from 23% in 2024), deploy on acquisition and conversion first, across hundreds of locations. Not because they love technology. Because they did the math.

So when the practice across the street captures every call, presents treatment in a way patients actually understand, and runs a profitable schedule, they're not "using AI better" than you.

They're taking your patients. And them using AI tools or not is not the problem: is the patient’s experience that matter, no matter how it is done.

Over the next editions, we'll take the revenue side apart one lever at a time:

Each time with the honest answer to the same question: where does AI genuinely help, and where is it just an expensive way to avoid a management problem?

Don't cancel your admin tools. We just want to nuance the whole thing and not confuse a cost-saving tool with a growth strategy.

One thing before you go.

You know your practice better than I do. So tell me where it hurts: hit reply to this email with one word, the lever you want me to break down first.

I read every reply, and the next edition goes where the most replies point. If your situation doesn't fit in one word, write me the full story. Those replies are usually the ones that end up shaping these emails.

Salim, co-founder of DentAI