Timeline awareness is not enough; compliance needs milestone-driven execution. The EU AI Act timeline should be translated into deliverables, owners, and review checkpoints so teams can show measurable readiness before each phase.
Milestone structure and implementation impact (Regulation (EU) 2024/1689)
1 August 2024 — Entry into force
The AI Act entered into force 20 days after publication in the Official Journal. This was the point to start role mapping (provider/deployer), system inventorying, and legal scoping.
2 February 2025 — First obligations apply
The prohibitions in Article 5 began to apply, and AI literacy expectations under Article 4 also became relevant. Organizations should already have no-go screening for prohibited practices and baseline staff literacy plans.
2 August 2025 — GPAI and governance layer
Core obligations for general-purpose AI models and governance architecture started applying. SMEs that depend on upstream GPAI providers should document model dependency controls and downstream safeguards.
2 August 2026 — Main compliance milestone
Most obligations for high-risk AI systems listed in Annex III become operationally critical by this date, including risk management, data governance, documentation, oversight, logging, and post-market processes.
2 August 2027 — Final phase for certain transitional elements
The final transitional layer applies for specific cases, including some legacy systems and sectoral interactions. By this point, regulators and enterprise customers expect mature evidence, not roadmap promises.
Deadline-to-deliverable planning model
T-12 to T-9 months
- finalize inventory completeness
- classify top-impact workflows
- identify unresolved legal ambiguities
T-9 to T-6 months
- implement oversight and logging in high-impact pathways
- launch incident and corrective-action governance
T-6 to T-3 months
- complete documentation bundle and evidence index
- run internal audit simulation
T-3 to T-0 months
- close critical residual gaps
- prepare external-facing compliance response pack
KPI dashboard for timeline execution
- % workflows with confirmed role/risk classification
- % high-impact workflows with documented oversight
- % controls with named owner and due date
- open critical issues and mean closure time
- evidence freshness score (last-reviewed age)
Final takeaway
Treat each AI Act date as a release milestone with acceptance criteria. Organizations that run timeline execution like product delivery avoid deadline panic and strengthen trust with customers, partners, and regulators.