The classification wizard guides business operators through a structured intake — capturing the business context, deployment scope, and technical indicators that the EU AI Act requires to be documented before risk classification can begin.
The wizard produces a complete, consistent system profile. Not a free-text description. A structured record that the governance chain builds on from registration through audit.
The classification wizard is the entry point for every AI system in EAB. It guides the business operator through a structured intake — covering the information the EU AI Act requires to be documented before risk classification can be conducted. The wizard does not accept free-text answers in place of structured responses. Every field has a defined schema. Every registration produces a comparable entry.
As the operator progresses through the wizard, EAB provides AI-assisted feedback — identifying potential risk indicators based on the answers given, flagging deployment contexts that are typically associated with Annex III classification, and surfacing questions about affected populations that operators sometimes miss. The AI analysis is a prompt, not a conclusion — risk classification happens in the supervised screening session, not in the wizard.
On completion, the wizard produces a business-layer system profile: anchored to the operator who created it, timestamped, and in a state that signals readiness for technical completion by the AI system owner. The technical completion stage adds the layer that the operator typically cannot provide — training data characteristics, model architecture, human oversight provisions, and Annex IV documentation indicators.
The wizard is the start of the governance chain — what it captures determines the quality of everything downstream.
Every registration field has a defined schema. System purpose is captured from a structured set of categories. Deployment context is selected from a defined list. The output is consistent across every system registered — by any operator, in any team.
Based on the answers provided, EAB identifies potential risk indicators — Annex III deployment contexts, affected population characteristics, decision influence patterns. These are surfaced to the operator as prompts — not as conclusions. Classification happens in the screening session.
Deployment contexts that are typically high-risk under Annex III — employee monitoring, credit scoring, educational assessment, law enforcement — are flagged immediately when selected. The operator is made aware of the classification implications before the technical profile is even started.
The wizard captures the information needed to make an initial actor role determination under Art. 2 and 3 — who developed the system, who controls it, and how it was placed on the market. The determination is confirmed by the supervisor during screening.
On completion, the wizard transitions the system to the technical completion stage and notifies the AI system owner. The business profile is locked — the technical owner adds to it, not over it. The operator retains visibility throughout.
Every completed registration produces an entry in the AI system registry — comparable, searchable, and ready for the governance chain. No two systems are registered in incompatible formats. The registry becomes a real inventory, not a list of descriptions.
Structured intake, AI-assisted risk indicators, consistent schema — the foundation of every compliance record in EAB.
EU-hosted · Anchored to CELEX 32024R1689
Tell us about your organization and what you’re looking to address. We’ll follow up with the relevant information.
