AI in medicine: what can be handed over and what never can

AI in medicine: what can be handed over and what never can

The conversation here starts not with possibilities but with a boundary: diagnosis, prescriptions and patients' personal data are never handed over under any circumstances, regardless of how good the model is. Everything else is the work around medicine, and more money is lost there than it appears.

The conversation here starts not with possibilities but with a boundary: diagnosis, prescriptions and patients' personal data are never handed over under any circumstances, regardless of how good the model is. Everything else is the work around medicine, and more money is lost there than it appears.

The main point first

Among those who answered my survey there are doctors and dentists. The subject is a live one for them, but the conversation has to start with the boundary rather than the possibilities.

What is never handed over under any circumstances:

Diagnosis. A model does not make a diagnosis, does not assess a patient's condition and does not confirm somebody else's diagnosis. An answer that looks plausible is more dangerous here than no answer.

Prescriptions. Dosages, regimes, interactions. The error is invisible from the look of the result: a figure looks like a figure and a drug name like a drug name.

Patients' personal data. Not sent out. Not in anonymised form "just in case", not in fragments.

Anything you sign for professionally. The decision is made by a specialist, and responsibility is not shared with a tool.

That boundary does not soften over time and does not depend on how good the model is. Everything below is about the work around medicine rather than about medicine.

Where the tool genuinely helps

The common feature: you supply the material, and it concerns a process rather than a patient.

Documents and procedures

Internal instructions, appointment protocols, preparation checklists for procedures, guidance for staff.

Analysing existing documents for contradictions and omissions is the safest and most useful task: the model compares what you gave it against itself.

Explaining the complex in plain language

One of a doctor's most frequent and least favourite tasks: explaining to a patient what is happening without jargon.

What matters here: you supply the content, the model works on the form. Not "explain what an implant is" but "rewrite this text of mine so a person with no medical training understands it, adding nothing".

Patient guidance

How to prepare for a procedure, what to do afterwards, when to seek help. Standard texts that in a clinic are either unwritten or written in the language of an equipment manual.

And the same principle: the content is yours, the form is the model's work. Every statement is verified by you before it goes to a patient.

Answers to standard questions

Not medical ones — organisational. Opening hours, cost, how to book, what to bring, how to get there.

That is what eats a receptionist's time and requires no medical qualification whatever.

Transcription and structuring

Dictated notes after an appointment, meeting recordings, going through patient feedback in bulk.

Administrative routine

Reporting, planning, texts for the site, handling reviews.

Separately: why this concerns the whole clinic rather than the doctor

The largest loss in a medical business is not clinical.

Enquiries outside working hours. Someone wrote in the evening, the answer came in the morning — by then he has booked somewhere else. That loss appears in no report: an enquiry left unanswered exists nowhere.

Expensive services. Asked "how much does an implant cost", a person gets a figure. A figure with no explanation of what it consists of and what he gets for it frightens people off — especially on large sums, where the decision takes time.

Patients who did not come back. Work with the existing base is usually not done: once a year the owner remembers and asks somebody to ring round, and it fizzles out.

Social media with no link to booking. Content goes out, but the route from a post to an appointment is not built.

All four are organisational tasks, and they are solved without a single medical statement.

A real case

The client is Zakhar Fomin, owner of a dental clinic. He agreed to be named along with the figures.

The task: a system that answers standard questions around the clock, qualifies the enquiry and books appointments. Effectively a smart receptionist.

Sold for $1 900 a year, paid in full. A separate follow-up order was a CRM at about $3 200. This is an example of one order, not a promise of income.

Why he agreed. The comparison was not against zero but against an existing cost line: an in-house team — a marketer, a social media specialist, a remote administrator — at around $720 a month plus taxes, holidays and the human factor.

And his main objection, stated outright: the fear that a bot would damage the reputation and lose an expensive implant patient.

How that is closed. Not by promising the bot is clever but by the architecture: it recognises the status of an enquiry, qualifies tactfully and hands a hot request to a live person. A large fee is not left on an automated exchange to the end.

That is the rule which matters more in medicine than anywhere: hot leads go to a human.

The verification protocol

For those who do use the tool on texts around the clinic.

Separate the verifiable from the reasoning. Names, dosages, timescales, standards, references are checked against the primary source always, even when they look obviously right.

Set the material frame. "Rely only on the attached text; if the answer is not in it, say the data is insufficient." Without that line the model will fill the gap with something plausible.

Nothing goes to a patient without a specialist's check.

And the safest mode: ask not "tell me about" but "rewrite this more clearly, adding nothing". When all the material is yours, there is nothing to invent.

How this is monetised in medicine

For a doctor. Time freed from administrative routine. Plus guidance and explanations that improve adherence to recommendations.

For someone working with clinics. This is one of the most solvent niches, and entering it requires no medical knowledge — because what is bought is not medicine but organisation: handling enquiries, working the patient base, content, analytics.

How to approach a clinic. Not with an offer of automation. With a question: how many enquiries arrive in the evening and at weekends, and what happens to them. The owner almost never knows the exact answer, and that not-knowing is your sale.

The order of prices. From our bot's feed of 5 986 jobs over four months: a system or a pipeline $490–1 460, against a median one-off job of around $60. And the $1 900 in the example above is the price of one specific project, not market statistics.

What the platform offers for medicine

Over 200 AI assistants across 14 categories. Relevant for a clinic: client retention, sales, marketing, HR, legal.

The legal department — thirteen assistants on Russian law, including the corporate lawyer and the legal assistant. They suit drafts of internal documents and a first review; your own lawyer signs off the result.

Gen AI, "File → text" — transcription and recognition. $0.15 per operation.

The express course "Perplexity as an investigative tool" — a method of verification. For a profession where an error costs your reputation it matters more than any other. $9, included in Basic.

n8n and Make automation templates — if a process is needed: handling enquiries, working the base, regular reminders. Standard templates come with the Creator plan.

Registration is free and opens three days of full Basic access.

Where to start today

One action, and it is not about medicine.

Take one piece of patient guidance you already have and ask for it to be rewritten more clearly — adding nothing and removing nothing in substance.

Read the result as a doctor and verify every statement.

That is the safest task there is: the content is yours, the checking is yours, and the model's work is only the form. That is where people begin.

The second approach: not a chat but a conveyor

Everything above is work in dialogue with an assistant. There is another format, and for volume it suits better.

Cascading prompts in spreadsheets

The “Earning with ready templates” course contains spreadsheets with an integrated model: a large task broken into micro-steps, each living in its own cell and taking the previous result as input. The formula can be dragged down a thousand rows — and a thousand tasks are processed at once. How this works step by step is covered separately: Thirty posts in one launch.

What is available for medicine:

The audience analysis template — patient segments, their fears and objections. The HR template — job descriptions and procedures for clinic staff. Neither touches diagnosis.

The course is included in the Basic plan.

The same chains, but as automation scenarios

When a chain needs calls to external services, a spreadsheet is not enough — there are n8n and Make templates for that: the same cascade, only as an automatic scenario where each step's result becomes the next one's input. Standard templates come with the Creator plan.

Why there is no list of ready assistants here

The catalogue holds over 200 roles across fourteen directions — marketing, law, HR, sales, engineering. There are no medical ones among them, and that is deliberate: responsibility for a diagnosis and a prescription is not shared with a tool.

What is applicable lies in a different plane — not "make a diagnosis" but taking the paperwork off a doctor. For that it matters where the data is processed. The Open Source section holds solutions deployed on your own server: patient records do not leave the clinic's perimeter. The catalogue is open to everyone, the project links from the Creator plan.

What this looks like on a real project

The organisational part described above was assembled for a dental clinic. The work was delivered in eleven sections, and it shows how this is built.

The audience is split into four segments. The anxious — priority number one, because the barrier here is fear rather than price. The aesthetically driven — an entry ticket with a high fee and months of consideration. The functional, with jaw joint problems — an open niche where nobody has found the patient a cause. The ambassadors — a loyal core who bring their families.

A principle worth remembering: a segment is an entrance, not a cage. People migrate: a functional patient becomes an aesthetic one, a satisfied one becomes an ambassador. The bot recognises the segment from the first phrases and reclassifies on new signals.

The running offer — a scan as a gift. The patient leaves with the image, a photo record and a treatment plan explained, even if he decides to be treated elsewhere. That is not a discount but the removal of risk: diagnostics like that are normally paid for separately.

And the medical frame written into all the material: we do not promise an outcome — we promise a cause and a plan · connections between oral health and the body are presented as an observation rather than a promise · anything acute goes straight to an in-person doctor · online is a second opinion rather than a diagnosis · no invented deadlines.

The metric the content was judged by was reactions and shares rather than views. A telling figure: the phrase "it does not hurt", presented as a slogan, produces just 1.12 shares. What works is not a promise but a story.

What to take on the platform for this task

What to read next

Local models for sensitive data — the only permissible option where data requirements apply.

What AI agents are, in plain words — how a smart receptionist is built.

A deal end to end: from a listing to $1 900 paid — the whole path of selling a system to a clinic.

How to verify a fact with AI search — the obligatory protocol for this profession.

Best open LLMs 2026 — the full guide to open models.

Also on AI SKILLS