All articles Trust & privacy

Trust by Design: Using AI in a Practice That Handles Sensitive Information

Oct 10, 2026 · By Rosa GJ · 7 min read

When dentists, vets and lawyers hear about AI assistants, the first question is rarely “does it work?” It is “can I trust it with my clients’ information?” That is exactly the right question.

Practices that handle health records, pet medical histories or confidential legal matters carry a responsibility most businesses don’t. A careless tool can damage a reputation built over decades, and expose the practice to real legal risk.

The good news is that trustworthy AI is possible. It simply has to be designed for trust from the start, not patched afterwards. This article explains the rules that apply, the principles that matter, and the questions to ask before you let any AI tool speak to your clients.

Why caution is justified

Your clients share things with you they share with almost no one else. That trust is the foundation of your practice, and it raises the bar for any technology that touches it.

• Dental clinics handle health information, which European data protection law treats as a special category requiring extra protection.
• Law firms are bound by professional secrecy. Even the fact that someone consulted a lawyer can be confidential, especially in areas like immigration, family or criminal law.
• Veterinary practices hold less sensitive data about the animals themselves, but still manage owners’ personal details, payment information and sometimes difficult conversations about end-of-life care.

Generic chatbots built for online shops were not designed with any of this in mind. A tool that stores conversations in unknown locations, uses them to train its models, or gives confident answers to medical or legal questions is a risk, not a convenience.

The rules that apply

Two sets of European rules shape how a practice can use an AI assistant: the GDPR for personal data, and the EU AI Act for transparency.

The GDPR already applies to every conversation an assistant has with your clients. In practice that means having a lawful basis for processing, collecting only what you need, telling clients how their data is used, keeping it secure and signing a data processing agreement with any provider that handles it on your behalf. Health data, like that held by dental clinics, is a special category that needs even stronger safeguards.

The EU AI Act adds a simple but important duty. Since 2 August 2026, AI systems that interact directly with people, such as chatbots and voice assistants, must make clear that the person is dealing with an AI, unless it is obvious from the context (Baker Botts). A July 2026 amendment package delayed other parts of the Act, but left this transparency rule in place (Technology Org).

On top of these, lawyers and healthcare professionals follow their own professional codes on confidentiality. A good AI setup respects all three layers.

This article is general information, not legal advice. Your professional association or data protection adviser can confirm what applies to your practice.

Six principles of trustworthy AI

Trust is not a feature you add at the end. It comes from a handful of design choices made before the assistant ever speaks to a client:

1. 1Be open about being AI. The assistant introduces itself as an automated assistant from the first message, and never pretends to be a member of staff.
2. 2Collect the minimum. It asks only for what is needed to help, such as a name, a contact number and the reason for the visit, and never invites clients to share full medical histories or case details in a chat.
3. 3Never give professional advice. It organises, informs and books. Diagnosis, treatment and legal opinions stay with your professionals.
4. 4Always offer a human. Clients can reach a person whenever they ask, and anything sensitive, urgent or emotional is passed to your team.
5. 5Keep data in safe hands. Conversations are stored securely, with clear retention periods, access limited to the people who need it, and providers that process data in line with the GDPR.
6. 6Don’t let client data train someone else’s model. Your clients’ conversations should be used to serve them, not to improve a technology company’s products.

These principles protect your clients, but they also protect you. They make the assistant easier to explain, easier to defend and easier for your team to trust.

How to explain it to your clients

Clients are far more comfortable with AI when you tell them openly what it does and what it doesn’t. A short, honest introduction does most of the work.

Here is an example of an assistant’s first message for a law firm:

A few other simple steps help:

• Mention the assistant in your privacy policy, in plain language.
• Tell your reception team how it works, so they can answer questions with confidence.
• Invite feedback in the first weeks, and adjust the wording if clients find anything confusing.

Transparency is not just a legal box to tick. It shows clients that your practice uses technology the same way it handles everything else: carefully and with their interests first.

Technology that earns trust

The practices that benefit most from AI will not be the ones that adopt it fastest, but the ones that adopt it most carefully. When an assistant is open about what it is, collects only what it needs and always leaves room for a human, it strengthens the trust your clients already place in you.

At autonomi, we build AI assistants for dental clinics, veterinary practices and law firms with these principles built in from day one. Every assistant introduces itself honestly, follows your rules on what it can and cannot say, and hands over to your team whenever a person is needed.

Technology that helps you grow.