Agent readiness in Europe: when a customerโs AI assistant contacts you
How the Agent Readiness Standardโs checks on consent, data and disclosure line up with the rules European service teams already work under.
Version 1.0 ยท 18 November 2026 ยท 4 min read
The short answer
When a customerโs AI assistant contacts a European company, the questions that matter most are about protection and openness. Does the service check the customer is really behind the request? Does it ask the customer, not the assistant, before anything irreversible? Does it ask only for the data it needs, and say that it is AI? The Agent Readiness Standard (ARS) scores exactly these, in English, with German and French in review.
General information, not legal advice.
Why this matters for European teams
European service teams already work under rules built on a few shared ideas: process only the data you need, be clear with people about how their data is used, and tell people when they are dealing with a machine. A customerโs assistant does not change those ideas. It changes who is in the conversation.
When an assistant asks to cancel a contract, change a booking or see an order, the service is talking to software acting for a person. The person still owns the data and the decision. A service that treats the assistantโs word as the customerโs own consent, or hands account details to whoever asks, puts that person at risk.
How ARS lines up with the rules you know
| Principle | Where it comes from (in general terms) | What ARS checks |
|---|---|---|
| Ask only for what you need | Data minimisation under the GDPR and UK GDPR | Data minimisation: the service asks only for data the job needs, and nothing sensitive in open chat |
| Be clear about data use | Transparency duties under the GDPR and UK GDPR | Privacy stated: the service explains where data goes and how to reach a person or a data protection officer |
| Tell people it is AI | Transparency duties for AI systems that interact with people under the EU AI Act (Article 50) | AI disclosure: the service says it is AI up front, or as soon as it is asked |
| Act only on the customerโs decision | Consent and good-outcome duties, including the UK Consumer Duty for regulated firms | Consent before irreversible actions, customer verification, confirmation to the customer |
| Donโt leave customers stuck | Good-outcome and complaint-handling expectations | Route to a person, clear limits |
A good ARS score does not show that a service meets any law. It shows how the service behaved, on a stated date, when a declared AI assistant tested it.
A 12-point checklist for European service teams
Your help pages can be read by machines: llms.txt, FAQ structured data, or AI crawlers allowed on help pages.
A declared assistant can reach your chat without a login or a CAPTCHA for general questions.
Your chat says it is AI at the start, or as soon as it is asked.
Answers give exact values: dates with the month named, amounts with a currency, full links.
The same policy question gets the same answer every time.
Account details are shared only after the customer logs in or verifies.
Cancellations, refunds and changes need the customerโs own confirmation, never the assistantโs.
Confirmations go to the customerโs own email or phone on file.
The chat asks for no ID numbers, card details or dates of birth for general questions.
The chat explains where data goes and links your privacy notice.
A customer can always reach a person, and the person sees the conversation.
Your chat never asks for a password or one-time code.
How a test runs in Europe
A declared AI assistant, acting for a reserved test customer, works through 11 everyday jobs on your live web chat in English. German and French tests follow once their scripts pass native review. The test never pretends to be a real customer, never confirms anything irreversible and never asks to be put through to your staff. Read the full method.
Next steps
Questions people ask
Do AI chatbots have to disclose that they are AI in the EU?
The EU AI Act includes transparency duties for AI systems that interact with people, in Article 50. ARS checks whether a service says it is AI up front, or as soon as it is asked. Check with your legal team how the rules apply to your service.
How does GDPR data minimisation apply to customer service chatbots?
Data minimisation means collecting only the personal data a purpose needs. In a chat, that means not asking for ID numbers, card details or dates of birth for general questions. ARS scores this as one of the four dimensions in Protect the customer.
Can a customerโs AI assistant cancel a contract on their behalf?
Whether a company accepts that is its own decision. ARS checks that the service asks the customer, not the assistant, before anything irreversible, for example by sending a confirmation to the customerโs own email or phone, or asking them to log in.
About
The Agent Readiness Standard is published by OnClarity. OnClarity (www.onclarity.com) is an AI customer experience platform that helps teams resolve, score and understand every customer conversation. The standard is free to use and adapt, with attribution.
Text licensed under CC BY 4.0. Rubric file ars-1.0.json licensed under Apache 2.0.
