Customer-service chatbots
Identify whether customers are interacting directly with AI and preserve the disclosure timing, wording, placement, interface version and any applicable exception analysis.
DOES YOUR CHATBOT NEED
AN EU AI ACT DISCLOSURE?
AI chatbots and assistants are a practical Article 50 question for many businesses now. The key is not simply whether a bot exists, but whether natural persons interact directly with the AI system, what role the organization holds, whether an exception applies, what disclosure is implemented and whether the evidence still matches the deployed version.
BOUNDARYThis page is educational and operational guidance. It is not legal advice, certification, conformity assessment or regulatory approval.
CURRENT SOURCE ANCHOR
OPEN OFFICIAL EU SOURCE ↗European Commission · Article 50 transparency guidance
The Commission’s Article 50 guidance addresses transparency obligations for providers and deployers of relevant AI systems. The applicable Article 50 transparency rules began applying on 2 August 2026.
WHAT BUSINESSES NEED TO ESTABLISH
Do not start with a generic checklist.
Start with the actual system, role, use case and evidence boundary. The same regulation can produce different obligations for different actors and systems.
AI copilots and assistants
Do not assume an internal or embedded assistant is outside scope. Record who interacts with it, in what context, under whose authority and for what intended purpose.
White-label and vendor chatbots
If another provider supplies the model or application, separate provider and deployer responsibilities instead of treating the vendor contract as the compliance determination.
Changing models and interfaces
A new model, UI, deployment mode or customer journey can change the factual basis behind a prior transparency determination and trigger revalidation.
THE TA-14 OPERATING ROUTE
Turn the question into a governed record.
The goal is not merely to reach an answer. It is to preserve what facts, evidence, scope and limitations supported that answer at that time.
Identify the deployed chatbot
Record the product, model or service, version, interface, intended purpose, audience, geography and deployment owner.
Determine actor and interaction route
Establish whether the organization is acting as provider, deployer or another actor and whether natural persons directly interact with the AI system.
Preserve disclosure evidence
Keep the disclosure wording, screenshot or interface capture, timing, placement, exception analysis and version linkage as bounded evidence.
Revalidate after material change
Recheck the route when the model, vendor, interface, intended purpose, user population or official source state changes.
READY TO OPERATE · START PAID ACCESS NOW
Move from reading about the EU AI Act to maintaining the record.
Choose the smallest operating tier that fits the portfolio today. Upgrade when system count, team size or governance scope actually requires it.
Evidence Passport
$19/MOKeep a living system-level evidence record with obligations, gaps, versions and revalidation state.
START EVIDENCE PASSPORT →Compliance Workspace
$49/MOCoordinate evidence, owners, documentation, incidents and team compliance work in one governed workspace.
START COMPLIANCE WORKSPACE →Governance Pro
$99/MOOperate broader high-risk, GPAI, FRIA, post-market and material-change governance across a growing portfolio.
START GOVERNANCE PRO →Institution
$499/MORun institutional governance with expanded users, authority workflows, examiner rooms and portfolio reporting.
START INSTITUTION →Need an independent human readiness examination?A governed readiness review is separate from software access and starts at $750.
REQUEST READINESS REVIEW →Not ready to subscribe? Classify first.
Use the free classifier to establish the system, intended purpose, possible actor role, EU exposure and unresolved facts. When continuing evidence infrastructure is needed, paid access begins at $19 per month. Independent human readiness review remains a separate service.
COMMON QUESTIONS
Questions businesses are asking now.
Do all chatbots need the same EU AI Act disclosure?
No. The applicable route depends on the system, actor, interaction context and relevant exceptions. The safe operational approach is to preserve a system-specific determination and its evidence.
Is a small “AI” badge enough?
A badge may be part of an implementation, but the legal and evidentiary question is broader: what obligation applies, when the person is informed, how clearly, what exception may apply and whether the implementation matches the current system.
What if our chatbot uses a third-party model?
Third-party technology does not eliminate the need to determine your own role and obligations. Provider, deployer and downstream relationships should be recorded separately.
Can we start free?
Yes. Use the free classifier to establish the chatbot, intended purpose, actor role, EU exposure and unresolved facts before selecting a paid evidence workspace or requesting a readiness review.