Smart help that stays in your practice
Dictation instead of typing, answers from the manual, drafts for letters and reports, and scanned post that describes itself. All on your own practice's computers.
Dictation listens while you work and turns what you say into procedures, with code, tooth and surfaces. What it is unsure of is marked; what you do not tick is not booked.
Say what you did. Odontium turns it into procedures, with code, tooth and surfaces, which you accept with a tick.
Speak pocket depths and bleeding with F9 or the microphone, and keep your hands with the patient.
Dictation knows the practitioner by voice, so what someone else says does not count as a procedure.
Say "reminder" or "follow-up appointment" and it is ready as a suggestion. One click opens the window filled in.
For the denturist: the three group tariffs that best fit what you dictate at the intake.
In TSE Mobile, on the road too. The speech recognition runs on the phone itself; the draft goes to the workstation.
Ask Odontium answers a question about the software from its own manual, with the steps, the pictures and a reference to where it says so. No patient data goes in.
A question about the software, answered from the manual: with the steps, the pictures and suggested follow-up questions.
A draft report for the patient's notes from a dictation, with only what was dictated.
A draft letter, an explanation for the patient in plain language or another language, or a shorter version of a text.
Describe a letter and Odontium writes the template, with fields where the details go.
Say what goes on a Dymo label and the design is there. Patient details are only added when printing.
Odontium reads scanned post and rejected claims along with you and says in plain language what they say and what needs doing.
Odontium reads the first page of a scan and suggests a description.
For a scanned letter: the sender, a short summary and a task when something needs doing.
Explains rejected claims in plain language, and flags what is in the report but was not booked.
The language model and the speech recognition run on the practice's computers. Nothing goes to the internet.
Every AI feature is off by default. The practice switches it on, and a workstation can switch it off again for itself.
What the AI writes appears in a box marked "Draft, made by AI". Nothing is saved or sent until you click Accept.
A voiceprint needs consent per staff member and stays on the computer where it was recorded.
No. The language model and the speech recognition run on the practice's own computers. Nothing goes to the internet, not even to TSE's server.
Dictation works on every workstation, faster with a graphics card. The features with the language model, such as Ask Odontium and the writing aid, need a PC with at least 16 GB of memory.
No. Every AI feature is off by default. The administrator switches on what the practice wants, and a workstation, such as the front desk without a microphone, can switch a feature off for itself.
You are. A draft appears in a box headed "Draft, made by AI" and is only saved or sent when you click Accept. The claims aid never suggests a code by itself.
In a demo we show Dictation and Ask Odontium, on your own kind of treatments.
KIWA-certified for information security and quality management.
© 2026 TSE B.V. All rights reserved.