---
title: "EU AI Act for Clinics: AI Phone Receptionists and Triage"
description: "A clinic AI receptionist that books and reminds must say it is AI from August 2026. Emergency triage is high-risk from December 2027. Where the line is."
url: "https://ainora.lt/blog/eu-ai-act-clinics-healthcare-ai-phone-2026"
published: "2026-11-11"
author: "Justas Butkus"
---

# EU AI Act for Clinics: AI Phone Receptionists and Triage

> **TL;DR:** An AI receptionist that books, moves and cancels appointments, sends reminders and answers questions about opening hours, prices or directions is not a high-risk system under the EU AI Act. It must tell patients it is an AI, from 2 August 2026. The line is triage: AI used as an emergency healthcare patient triage system is high-risk under Annex III point 5(d), with full duties from 2 December 2027. AI that gives diagnosis or treatment advice may be a medical device, which is a separate route.

> **Not legal or medical advice**
>
> This is general information about Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744. Every legal point links to its source. All sources were last checked on 11 November 2026. Your own counsel and your medical lead decide how the rules apply to your practice.

- **2 Aug 2026** - AI receptionists must tell patients they are AI (Source: AI Act, Art. 50 - https://artificialintelligenceact.eu/article/50/)
- **5(d)** - Annex III point: emergency healthcare patient triage systems are high-risk (Source: AI Act, Annex III - https://artificialintelligenceact.eu/annex/3/)
- **2 Dec 2027** - Annex III high-risk duties apply, after the 2026 delay (Source: Reg. (EU) 2026/1744 - https://eur-lex.europa.eu/eli/reg/2026/1744/oj)
- **2 Aug 2028** - high-risk duties for AI in products such as medical devices (Annex I) (Source: AI Act, Art. 6(1) - https://artificialintelligenceact.eu/article/6/)

## Is an AI receptionist for a clinic high-risk under the EU AI Act?

No, not for front-desk work. The AI Act does not make healthcare a high-risk sector as a whole. It lists specific uses in [Annex III](https://artificialintelligenceact.eu/annex/3/), and the healthcare-related ones are narrow: emergency triage, emergency call handling and dispatch, and public authorities deciding who is eligible for public healthcare services.

A phone agent that books a check-up, moves a cleaning appointment or tells a caller where to park does none of those things. What it does is talk to people, so it falls under [Article 50](https://artificialintelligenceact.eu/article/50/): patients must be told they are talking to an AI.

## Which clinic phone tasks are high-risk, and which are not?

| What the AI does on the phone | Status under the AI Act | What it triggers | From |
| --- | --- | --- | --- |
| Books, moves and cancels appointments | Not high-risk | Tell patients it is an AI (Art. 50) | 2 Aug 2026 |
| Sends appointment reminders | Not high-risk | AI disclosure (Art. 50) | 2 Aug 2026 |
| Answers questions: hours, prices, directions, documents | Not high-risk | AI disclosure (Art. 50) | 2 Aug 2026 |
| Takes a prescription renewal request and passes it to staff | Not high-risk, if staff decide | AI disclosure (Art. 50) | 2 Aug 2026 |
| Hears a red-flag word and follows your fixed script: emergency number plus transfer | Designed to hand over, not to triage; confirm with counsel | AI disclosure; clinician-approved script | 2 Aug 2026 |
| Assesses symptoms to rank urgency in emergency care | High-risk (Annex III 5(d)) | Deployer duties (Art. 26) | 2 Dec 2027 |
| Gives diagnosis or treatment advice | May be a medical device (MDR); high-risk via Annex I if it needs a third-party assessment | Separate medical device route | 2 Aug 2028 under the AI Act |
| Decides eligibility for public healthcare, for a public authority | High-risk (Annex III 5(a)) | Deployer duties and impact assessment | 2 Dec 2027 |
| Recognizes patients’ emotions from their voice | High-risk (Annex III 1(c)) | Notice to patients (Art. 50(3)), deployer duties | 2 Aug 2026 notice, 2 Dec 2027 duties |

[Source: AI Act, Annex III and Art. 50](https://artificialintelligenceact.eu/annex/3/)

Point 5(a) is often misread. It covers AI "used by public authorities or on behalf of public authorities to evaluate the eligibility of natural persons for essential public assistance benefits and services, including healthcare services". A private clinic booking appointments is not evaluating anyone's eligibility for public benefits.

## Where does triage start, and why does it matter?

[Annex III point 5(d)](https://artificialintelligenceact.eu/annex/3/) lists AI systems "intended to evaluate and classify emergency calls by natural persons or to be used to dispatch, or to establish priority in the dispatching of, emergency first response services ... as well as of emergency healthcare patient triage systems". If your phone agent is meant to judge how urgent a patient is and decide who is seen first in emergency care, it is in this category.

A receptionist agent can be built so that it never makes that judgment. The clinic writes a short list of red-flag phrases with its clinicians, such as chest pain or trouble breathing. When a caller says one, the agent follows a fixed script: it tells the caller to call the emergency number and transfers the call to staff. It does not weigh symptoms, score urgency or decide who waits. That is a handover, not triage. Point 5(d) also covers AI that evaluates and classifies emergency calls, so ask your counsel to check that part as well. Whether your exact design stays on the right side is their call, and the script itself is a question for your medical lead.

## Could a clinic AI phone agent be a medical device?

Only if it does medical work. Software with a medical purpose, such as giving diagnosis or treatment advice, may qualify as a medical device under the EU Medical Device Regulation. That is a separate route into the AI Act: under [Article 6(1)](https://artificialintelligenceact.eu/article/6/), an AI system that is a product covered by the listed EU product laws, or a safety component of one, and that needs a third-party conformity assessment, is high-risk. After the 2026 amendment, those duties apply from 2 August 2028.

A receptionist that books, informs and hands over does not give medical advice. Keep it that way in the script: "I can't give medical advice, but I can book you with a doctor or put you through to the nurse."

## What must a clinic AI agent tell patients?

[Article 50(1)](https://artificialintelligenceact.eu/article/50/) requires AI systems that talk to people to be designed so that people know they are dealing with an AI, unless that is obvious. Article 50(5) says the information must be clear and given at the latest at the first interaction. On the phone, that is the greeting.

A clear opening: "Good morning, you've reached the AI assistant of [Clinic]. I can book, move or cancel your appointment, or put you through to reception." Use the same line for reminder calls, and answer honestly whenever a patient asks if they are speaking to a person. Breaching this duty can be fined up to EUR 15 million or 3% of worldwide annual turnover; for SMEs the lower of the two applies ([Article 99](https://artificialintelligenceact.eu/article/99/)).

## What are the key dates for clinics?

| Date | What applies | What it means for a clinic phone line |
| --- | --- | --- |
| 2 Feb 2025 | Prohibited practices (Art. 5), AI literacy (Art. 4) | Staff using AI tools get support to understand them; no AI tracking of staff emotions |
| 2 Aug 2026 | Transparency (Art. 50) | The AI receptionist says it is an AI |
| 2 Dec 2027 | Annex III high-risk duties, after the 2026 delay | Emergency triage AI follows Art. 26 |
| 2 Aug 2028 | High-risk duties for AI in Annex I products | Medical device AI that needs a third-party assessment |

[Source: Reg. (EU) 2026/1744](https://eur-lex.europa.eu/eli/reg/2026/1744/oj)

The ban on inferring emotions at the workplace ([Article 5(1)(f)](https://artificialintelligenceact.eu/article/5/)) applies to clinics as employers too: no AI that reads your receptionists' or nurses' emotions from their voice, unless it is for medical or safety reasons.

## What about patients' health data?

The AI Act sits on top of GDPR; it does not replace it. Patients mention symptoms, medication and conditions on the phone, and health data is a special category under GDPR. So the data side matters as much as the AI side: who can open the recordings and transcripts, how long you keep them, and what the agent writes into your practice system. Your DPO sets those rules. Our [GDPR guide for AI voice agents](/blog/ai-voice-agent-gdpr-compliance-guide) covers the questions to ask.

## How should a clinic set up its AI receptionist?

1. **Write the disclosure into the greeting** - The first sentence says it is an AI assistant and how to reach a person. (Source: AI Act, Art. 50 - https://artificialintelligenceact.eu/article/50/)
2. **Keep it to admin** - Booking, moving, cancelling, reminders and practical questions. No diagnosis, no treatment advice, no urgency scoring.
3. **Agree a red-flag script with your clinicians** - A short list of phrases that always lead to the same action: the emergency number and a transfer to staff. The agent does not weigh symptoms.
4. **Set clear handover rules** - Anyone who asks for a person, anyone upset or confused, and anything clinical goes to your team.
5. **Limit access to transcripts** - Health details end up in call records. Give access only to the people who need it, and set how long records are kept.
6. **Write down the classification** - One page: what the agent does, what it does not do, and why it is not triage. Update it when the flow changes.

## How does an AI phone agent from AInora fit?

AInora builds AI receptionists for clinics and practices: booking, rescheduling, reminders and answers to everyday questions. How it lines up with the points above:

- The agent says it is an AI assistant in its first sentence, in the words you choose.
- It hands the call to your team when the patient asks, or when your rules say so, including your red-flag script.
- Every call gets a transcript and a summary your staff can review.
- You decide what it may and may not say. No medical advice means no medical advice.
- It speaks many languages, so patients can book in the language they are most comfortable with.

Your counsel and your medical lead decide the rules; the agent follows them. To hear one with your own services and opening hours, [fill out the short form](/contact?from=blog-eu-ai-act-clinics-healthcare-ai-phone-2026) and we build a demo receptionist from your website. More on the use case: [AI answering service for medical practices](/blog/ai-answering-service-for-medical-practices).

## Frequently Asked Questions

**Is an AI receptionist for a medical practice high-risk under the EU AI Act?**

Not when it books appointments, sends reminders and answers practical questions. Annex III lists emergency healthcare patient triage and emergency call handling, plus public authorities deciding eligibility for public healthcare. An AI receptionist must still tell patients it is an AI from 2 August 2026. (Source: AI Act, Annex III - https://artificialintelligenceact.eu/annex/3/)

**Is AI triage high-risk?**

Yes, for emergency healthcare patient triage systems and AI that classifies emergency calls or sets dispatch priority (Annex III point 5(d)). The duties apply from 2 December 2027, after the delay in Regulation (EU) 2026/1744. (Source: AI Act, Annex III 5(d) - https://artificialintelligenceact.eu/annex/3/)

**Does Annex III 5(a) cover private clinics?**

No. Point 5(a) covers AI used by public authorities, or on their behalf, to evaluate eligibility for essential public assistance benefits and services, including healthcare services. It does not cover a private clinic’s booking line. (Source: AI Act, Annex III 5(a) - https://artificialintelligenceact.eu/annex/3/)

**Does a dental or physio clinic need to tell patients the receptionist is AI?**

Yes. From 2 August 2026, Article 50 requires that people talking to an AI system are told so, clearly and at the latest at the first interaction, unless it is obvious. On a phone call it is not obvious. (Source: AI Act, Art. 50 - https://artificialintelligenceact.eu/article/50/)

**Can an AI receptionist give medical advice?**

It should not. Software that gives diagnosis or treatment advice may qualify as a medical device, and AI that is a medical device needing a third-party conformity assessment is high-risk under Article 6(1) from 2 August 2028. A receptionist should book, inform and hand over. (Source: AI Act, Art. 6(1) - https://artificialintelligenceact.eu/article/6/)

**What should the agent do if a caller describes an emergency?**

Follow a fixed script your clinicians approve: tell the caller to call the emergency number and transfer to staff. The agent should not assess symptoms or decide urgency itself.

**Does GDPR still apply to AI receptionist calls?**

Yes. The AI Act adds rules for the AI system, and GDPR still governs the personal data in each call. Health details are a special category under GDPR, so access to recordings and transcripts should be limited.

Related on AInora

Explore the platform and industry pages relevant to this article.

- [AI for medical clinicsPatient scheduling and triage](/industries/medical-clinics)
- [AINORA AI voice agentPlatform overview and capabilities](/ai-voice-agent)
- [AI debt collectionCompliant voice for recoveries](/ai-debt-collection)
- [PricingPriced individually, after a call](/pricing)
- [How it worksSetup, integrations, and go-live](/how-it-works)