Back to Knowledge Hub
Security & Compliance

EU AI Act Compliance: Risk Classification, Ownership & Human-in-the-Loop

What "EU AI Act ready" actually means. Risk classification per use case, department owners for accountability, and built-in safeguards.

What the EU AI Act requires

The EU AI Act classifies AI systems by risk level and imposes obligations accordingly. For enterprise AI assistants, the key requirements are: risk assessment per use case, human oversight for high-risk applications, transparency about AI-generated content, and documentation for accountability.

“EU AI Act ready” is not a checkbox. It’s a set of capabilities built into the platform architecture. Here’s how the Plainsight AI Assistant supports compliance across different risk levels.

Risk classification by department

Different departments use AI differently, and they face different compliance obligations. Here’s how risk classification maps to typical enterprise use cases:

HRHigh Risk

AI systems used in employment decisions (screening, evaluation, promotion) are classified as high-risk under the EU AI Act. Requires documented risk assessment, human oversight, and transparency obligations.

Documented risk classification in admin panel
Department owner designated as accountable person
Human review required before any HR decision
Full audit trail with source citations
Customer ServiceLimited Risk

AI chatbots and assistants interacting with customers have transparency requirements. Users must know they are interacting with AI. Lower compliance burden than high-risk systems.

Disclosure that responses are AI-generated
Source citations on all answers
Escalation path to human agents
Conversation logs for quality review
LegalLimited Risk

AI used for document analysis, contract review, and legal research. Not making autonomous decisions, supporting human professionals. Transparency and accuracy are key.

Source citations with document links
Clear indication of AI confidence levels
Human-in-the-loop for all outputs
Audit trail per conversation
FinanceMinimal Risk

AI used for data analysis, chart generation, and report drafting. The AI processes data but does not make financial decisions. Minimal compliance obligations.

Data analysis runs in sandboxed environment
Results are verified by finance professionals
No autonomous financial decisions
Usage tracked in admin dashboard

How the Plainsight AI Assistant supports compliance

Risk classification per use case

Each department environment can be tagged with its EU AI Act risk level. This classification is documented in the admin panel and included in compliance exports.

Department owner designation

Every department has a designated owner, the person accountable for how AI is used in that context. Owners receive usage reports and can configure safeguards.

Human-in-the-loop

For high-risk use cases, the platform supports human review workflows. AI outputs are flagged for human approval before action is taken. No autonomous decisions in sensitive contexts.

Source citations & audit trail

Every AI response includes citations linking back to source documents. Full conversation history is retained for audit. Compliance teams can export audit logs at any time.

Data protection by design

Department isolation ensures data boundaries are enforced architecturally, not just by policy. Combined with GDPR-compliant data residency, this satisfies the Act's data governance requirements.

Ready to see it in action?

Schedule a personalised demo and see how the Plainsight AI Assistant fits your organisation.

Request a demo