AI Regulatory Navigator

Describe your AI system in plain language. The Navigator — an advanced graph RAG system — classifies it against 1,000+ high-risk AI system archetypes and 81 regulatory instruments to produce a multi-framework risk classification with cited regulatory references. No hallucinated Articles. No invented obligations. Every recommendation traceable to its source.

Architecture

Graph RAG — not just another chatbot

The Navigator isn't a language model with regulatory documents in its context window. It's a structured knowledge graph with retrieval-augmented generation — traversing relationships between regulations, archetypes, and obligations.

Knowledge Graph Architecture

Graph RAG

The Navigator is built on a dense knowledge graph connecting regulatory text, risk archetypes, compliance obligations, and precedent classifications. Unlike flat document retrieval, graph traversal follows the relationships between regulations — how an Annex III category interacts with GDPR special category data, how NIS2 cybersecurity requirements compound with AI Act obligations for the same system.

1,000+ High-Risk Archetypes

Classification

The graph contains over 1,000 classified AI system archetypes — real-world system profiles spanning credit scoring, medical diagnostics, autonomous vehicles, recruitment filtering, predictive policing, content moderation, and hundreds more. Your system description is matched against these archetypes to identify the closest regulatory profile, reducing classification ambiguity.

81 Regulatory Instruments

Multi-framework

The Navigator monitors 81 regulatory instruments across the EU AI Act, GDPR, NIS2, sector-specific regulations (MDR, MiFID II, DORA, Solvency II), national implementations, and voluntary standards (ISO 42001, ISO 27001, NIST AI RMF). A single query returns obligations from every applicable instrument — not just the one you asked about.

Two-Stage Citation Validation

Verified

Every regulatory reference in the Navigator's response is validated twice — first against the source text to confirm the citation exists, then against the context to confirm it says what the response claims. Fabricated references are suppressed before they reach the user. No hallucinated Articles, no invented obligations.

Classification Process

From description to classification in five steps

A structured process that takes your plain-language system description and produces an actionable, multi-framework risk classification.

Describe Your AI System

In plain language, describe what your AI system does, what data it processes, who uses it, who is affected by its outputs, where it operates, and what decisions it influences. The Navigator accepts natural language — you don't need to know regulatory terminology to get started.

Graph Traversal & Archetype Matching

Your description is decomposed into regulatory dimensions — intended purpose, data categories, deployment sector, affected persons, deployment geography, and supply chain position. The graph traverses relationships between these dimensions to identify applicable instruments and match against the 1,000+ archetype library.

Multi-Framework Risk Classification

The Navigator produces a preliminary risk classification under each applicable framework. EU AI Act risk tier (prohibited, high-risk, limited, minimal), GDPR processing category, NIS2 entity classification, and sector-specific classifications. Each classification is cited to the specific Article and criteria that apply.

Obligation Decomposition

From the classification, the Navigator decomposes the specific obligations your system triggers — documentation requirements, testing obligations, monitoring mandates, reporting duties. Each obligation is traceable to its regulatory source and cross-referenced against obligations from other applicable instruments.

Guidance & Recommendations

Actionable recommendations for your next steps — which TRACE phase to enter, which documentation to prepare, which assessments to conduct. The Navigator identifies the most efficient path through your compliance obligations based on your system's specific profile.

81 Instruments

The regulatory corpus

The Navigator doesn't just know the EU AI Act. It monitors 81 regulatory instruments across six categories — and understands how they interact.

Navigator

Data Schema Analysis

Combined View

Combined Intelligence

Navigator + Data Schema Analysis

Classification tells you what regulations apply. Data governance tells you what your data looks like. Together, they provide a comprehensive view of your AI system's risk posture.

Navigator

Risk tier, applicable instruments, obligation map

Classifies your AI system against regulatory frameworks

Data Schema Analysis

Data classification, sensitivity flags, lineage gaps

Maps your data architecture against governance requirements

Combined View

Unified classification + data governance report for decision-making

Comprehensive risk posture assessment