ENFORCEMENT COUNTDOWN ACTIVE

The EU AI Act Becomes
Enforceable August 2, 2026

Most organizations deploying AI in regulated industries cannot demonstrate where their training data came from, what reasoning produced their outputs, or who authorized the decisions their systems make. In 103 days, that becomes a legal liability.

Regulation (EU) 2024/1689  ·  103 days remaining

Can your AI answer these when a regulator asks?

The EU AI Act requires high-risk AI systems to satisfy specific, enforceable requirements. These four questions distill Articles 9 through 15 into the architectural test that determines compliance.

01
Where did the training data come from?
Article 10 — Data & Data Governance
02
What reasoning produced this output?
Articles 11 & 12 — Documentation & Record-Keeping
03
Who authorized this decision?
Article 14 — Human Oversight
04
Can you reconstruct the chain if a regulator asks?
Articles 9 & 13 — Risk Management & Transparency
If your architecture can't answer these, compliance won't save you.

The EU AI Act doesn't care where you're headquartered

A SaaS vendor in California screening résumés for a French employer must comply. A financial platform in London serving German clients must comply. A legal tech company in Toronto processing EU case data must comply.

From exposure to compliance — before August 2

Calyx Intelligence provides a complete EU AI Act compliance pathway. Not a checklist. Not a policy review. An architecture-level evaluation and remediation plan that produces the evidence regulators will actually ask for.

Governance Readiness Assessment
Fixed-scope engagement  ·  3–4 weeks
Map your current AI governance posture against EU AI Act Articles 9–15. Identify gaps in data provenance, audit trail infrastructure, human oversight, and technical documentation. Deliver a clear picture of where you stand and where you're exposed.
Available now
Remediation Roadmap
Scoped to assessment findings  ·  4–6 weeks
Design the governance architecture that closes the gaps identified in Phase 1. Authority chains, audit trail infrastructure, human oversight protocols, data provenance frameworks. A specific, implementable blueprint — not a generic compliance template.
Available now
Compliance Documentation
Scoped to remediation roadmap  ·  4–8 weeks
Produce the artifacts regulators will review: technical documentation (Article 11), risk management documentation (Article 9), data governance records (Article 10), conformity assessment preparation (Article 43), and EU database registration materials (Article 49).
Available now
Architecture Implementation
Scoped per engagement
Deploy governance infrastructure into your AI operations. Integrate provenance logging, human authority gates, decision-chain documentation, and compliance gating into your existing systems. Model-agnostic — works regardless of which AI provider runs underneath.
Phase 2 engagement
Ongoing Governance
Scoped per client
Continuous compliance monitoring as the EU AI Act evolves and your AI deployments expand. Ongoing audit readiness, governance reporting, regulatory change tracking, and architecture updates. The recurring layer that keeps you compliant — not just today, but as the regulations tighten.
Post-implementation

The penalties are not theoretical

€35M
or 7% of global annual turnover for prohibited AI practices
€15M
or 3% of global annual turnover for high-risk system non-compliance

These are enforceable fines under Regulation (EU) 2024/1689. The higher amount applies. For a company with €1 billion in annual revenue, a 3% penalty is €30 million. The cost of a governance readiness assessment is a rounding error by comparison.

Governance-first architecture, not bolt-on compliance

Most AI platforms treat governance as a feature added after the system is built. Calyx Intelligence was designed from the foundation as decision infrastructure for regulated environments. The governance isn't layered on — it's the spine.

Model-agnostic by design. The governance architecture operates above the model layer. Whether you run OpenAI, Anthropic, Google, Mistral, or open-source models, the audit trail, human oversight, and provenance chain remain intact. Vendor lock-in is a flag item, not a feature.

Multi-vertical, multi-jurisdictional. The same governance stack that operates in legal (Juris), financial services (Numera), and healthcare (Clinica) verticals works across any regulated industry and any jurisdiction. Build to the EU standard. Satisfy every lesser standard by default.

You can't bolt provenance onto a system that wasn't built to produce it. You can only reconstruct it after the fact — which is forensics, not governance.

103 days until enforcement.
Start with a conversation.

The governance readiness assessment is the entry point. Fixed scope, clear deliverable, architecture-level evaluation. Everything that follows builds on what we find.

Start a Conversation
Or reach out directly: info@calyxos.ai