The documentation that follows you home. Factor justifications written from scratch at nine in the evening. Recall letters nobody gets round to. AI can take that work — but not through a consumer chat window: findings and treatment notes are special-category data under Art. 9 GDPR, with professional secrecy under § 203 StGB on top. So we do the opposite of the usual advice. Instead of sending your data to the model, we look at putting the model where your data already is.
Every practice knows the shape of it. Documentation moves into the evening because the law wants it prompt and complete — anamnesis, findings, duration, materials, what the patient was told and consented to. For private billing it tightens further, since fee items with a time requirement need the time written down.
Meanwhile a vacant assistant post takes months to fill, and on the income side there is little to turn: statutory fees are budgeted and the points are set. The lever a practice genuinely controls is its hours and its costs.
That is where AI earns its keep — provided the confidential part of the work never ends up somewhere it should not be.
Not every office needs all of these. Which ones are worth building depends on your workflows, your systems and what the hours actually cost you.
Shorthand or a dictated sentence becomes a structured, complete entry in your own wording, ready to be taken into the record. Prompt and complete becomes the fast option instead of the evening one.
A long history condensed to what matters before the patient is in the chair: previous findings, medication, allergies, what was planned last time and never happened.
Factor justifications that read as reasoning rather than a recycled block, plan accompanying letters, and the correspondence when an insurer queries a factor.
Reply drafts, explanation texts a patient actually understands, recall letters that go out on schedule instead of when someone finds an hour.
Quality-management manual, billing rules, hygiene plans and the things only your longest-serving assistant knows — searchable, answering in seconds.
Diagnose. Decide treatment. Replace the person holding the handpiece. It drafts; a human reads and approves every text before it goes anywhere.
Three routes are normally on the table, and the right one is decided per task rather than per company.
If tools already cleared by your organisation can do the job, using what you have is usually the shortest route and the easiest to defend.
For work that does not touch confidential material, a European service under a proper processing agreement is often the proportionate answer.
Where genuinely confidential material is processed, keeping the system inside the building can be the more sensible option. It is a possible answer, not automatically the right one.
Nothing gets built before the numbers are on the table. The first step is short, deliberately bounded, and useful even if you stop there.
One to two weeks. Your workflows mapped, the real effort quantified, the data classes sorted, your systems examined for what they actually allow — ending in a prioritised list and a plan for the first thing worth doing. The report is yours to keep.
One use case taken to production against acceptance criteria agreed in advance, measured against the baseline from the check, and handed over with the training to run it.
Keeping it current and monitored once it is live, with a periodic review of whether it is still earning its place — and the training record the EU AI Act expects of anyone whose staff use AI.
Scope, timing and a fixed price are agreed in writing before any work begins.
An honest filter costs both of us less than a wasted first meeting.
No. A good deal of what costs you time — appointment organisation, callbacks, reminders — involves no treatment data at all. Where treatment data is genuinely involved, the check establishes which architecture is defensible before anything is set up.
That depends on the system and your licence. The check looks at which exports or interfaces actually exist. Nothing about integration is promised before it has been verified.
Not necessarily. Whether additional equipment is needed depends on the use case and the architecture chosen. That question is answered before anything is bought.
You are, exactly as with a text drafted by an assistant. Which is why nothing is set up to send or file by itself — every text stays a draft until a person approves it.
With your figures, not industry averages: the actual hours in the affected workflows against the expected effort of the solution. If it does not add up for a practice your size, the report says so.
Three aggregated numbers are enough to start: how many appointments you run, how many are missed or cancelled at short notice, and roughly how many hours a week go into the phone and documentation. From those we can tell whether this is worth your time at all.
Request a callFree and without obligation · direct with the founder
