I remember the day I first sat down and told myself: let's see what this can really do in my practice — not in polished demo clips, but in an ordinary working day. Here's what that first week looked like, without the gloss: what worked, what surprised me, and where I had to be careful.
Day 1: a message to a patient in the right tone
I started with something small. I had to cancel and reschedule an appointment, and I wanted the message to sound neither curt nor overly apologetic. I asked AI to help me word it — and in seconds I had three versions, one of which I tweaked a little and sent. I was surprised how natural it sounded. I also learned my first rule: always check the name, date, and time before sending. I later organized ready-made examples of messages like this into the AI assistant for your practice.
Day 2: an official email without the struggle
On the second day I had to write a formal email to a supplier about a complaint. That's the kind of writing that usually eats half an hour because I weigh every word. AI gave me a solid draft in two minutes — formal, clear, without unnecessary sharpness. I added the specific details and sent it.
Day 3: a photo of a device
I had an older unit whose model label I couldn't find. I took a photo and asked AI what it was and its rough specs. It wasn't infallible, but it pointed me far enough to quickly find the right manual. A small trick that saved me a lot of digging.
Day 4: comparing before a purchase
Ahead of a bigger purchase, I compared a few units — I asked it to summarize the differences into a table, based on what I gave it. I made the decision myself, of course, but I got to an overview far faster than reading everything by hand.
Day 5: when I caught it making a mistake
And then — the most important lesson of the whole week. I asked it something technical and it confidently gave me an answer that sounded right but wasn't. That's called a hallucination, and I later wrote about it in more detail in the post on the limits of artificial intelligence. Since then I verify anything important from a second source. And one rule I never break: I don't enter patient-identifying data into public AI tools — why and how, I explained in the post on patient privacy.
What I concluded after a week
It's not magic, and it's certainly not a replacement for the dentist. But it is a quiet assistant that takes over the tedious, repetitive parts of the day and gives me back time for what only I can do — working with the patient. What surprised me most wasn't "how smart it is," but how quickly it fit into perfectly ordinary little tasks.
If you're just starting out too
You don't have to begin with the big things. Open ChatGPT, give it one boring task that's on your desk today, and see what comes back. If you want a step-by-step order, there's the guide on getting started with AI, and to make AI do exactly what you ask, the post on dental prompts helps — or try the prompt generator right away.
Test your knowledge with the short AI quiz, and for new tools and tips I try first-hand subscribe to the newsletter — once or twice a month, no spam.
