Skip to content
Sencai.docs

EU AI Act compliance

The EU AI Act (Regulation 2024/1689) requires organizations deploying AI systems to classify them by risk and, depending on the category, meet specific conformity and registration obligations. Sencai’s EU AI Act screen is a review queue for tracking that classification work for the AI systems your organization runs or deploys.

Find it at Compliance & Audit → EU AI Act in the app (/gravity/compliance/ai-act).

What this screen shows

Section titled “What this screen shows”

Each row is one AI system your organization has recorded, with:

  • System name and a short use-case description
  • Risk category - Unacceptable, High, Limited, or Minimal, per the Act’s own risk tiers
  • Review status - Draft, Under Review, Approved, Rejected, or Archived
  • Conformity assessment - Not Required, Self Assessment, or Third Party
  • Whether registration in an EU database is required for that system
  • The next review date, with an overdue indicator if it’s passed

Summary cards at the top show the total number of AI systems tracked, how many fall in the High or Unacceptable categories, and how many reviews are still pending. Filter chips let you narrow the table to a single risk category.

The list is scoped to your organization - you see only the AI systems recorded for the organization you’re currently working in.

What the risk categories mean

Section titled “What the risk categories mean”

The four categories on this screen follow the Act’s own risk tiers, not a Sencai-specific scale:

CategoryWhat it covers
UnacceptablePractices the Act prohibits outright (for example certain forms of manipulative or biometric-categorization systems)
HighSystems used in contexts like employment, credit, or critical infrastructure, subject to the Act’s strictest conformity, documentation, and oversight requirements
LimitedSystems with specific transparency obligations (for example, users must be told they’re interacting with AI)
MinimalEverything else - the Act imposes no mandatory obligations, though voluntary codes of conduct exist

Which category a given AI system actually falls into is a legal classification your organization makes - Sencai doesn’t determine it for you.

What you can do today

Section titled “What you can do today”

Right now this screen is a review and tracking view: you can see every recorded AI system, its risk category, its review status, and filter by risk category. Adding a new AI system review from this screen isn’t available yet - the “New Review” action currently shows a “coming soon” message rather than opening a form. If you need to register a new AI system for review, contact your account team.

What Sencai records elsewhere that helps

Section titled “What Sencai records elsewhere that helps”

Two related, separate pieces of Sencai are relevant to your EU AI Act program even though they’re not on this screen:

  • AI governance settings (Intelligence → AI governance) track your organization’s configured autonomy level for Sencai’s own AI features (how much Sencai’s AI assistant is allowed to act without a human confirming each step) and Lumen’s own usage/rate limits - a different, narrower thing than classifying your own AI systems, but relevant if part of your AI Act inventory includes how you use Sencai’s built-in AI capabilities.
  • LLM usage records track token and cost usage for AI features across your organization, which can be a useful input if part of your risk assessment considers what AI capabilities are actually in active use.

What’s not here

Section titled “What’s not here”

This is not a full EU AI Act compliance workflow. There’s no built-in conformity-assessment generator, no automated risk-classification logic, and no direct integration with the EU’s AI system registration database. The screen is a structured place to record and track the classification work your organization does - the classification judgment itself, and any resulting conformity or registration filing, remains your organization’s responsibility.

What’s next

Section titled “What’s next”