EU Data Act Application Day: AI Platform Switching and Compliance

As of 12 September 2025, the EU Data Act is fully applicable, meaning EU customers can invoke Article 23 switching rights immediately. Organizations operating AI platforms must track export requests, verify that Article 25 fee schedules are published, and ensure smart-contract termination safeguards under Article 30 activate correctly when customers shift workloads. This application day marks a significant operational milestone requiring sustained compliance attention.

Application date significance and immediate obligations

The Data Act's application date triggers immediate enforceability of customer rights that organizations must be prepared to honor. Unlike many regulations that provide transition periods after application dates, the Data Act expects organizations to have compliance capabilities operational from day one. Customers can submit switching requests immediately, and organizations must respond within mandated timelines.

Article 23 establishes that customers have the right to switch data processing service providers without undue delay. Organizations must have documented procedures for receiving, processing, and fulfilling switching requests. Customer service teams need training on proper handling of switching requests and escalation procedures for complex scenarios.

The application date also triggers transparency obligations under Article 24 requiring publication of detailed information about data processing services, switching procedures, and associated fees. Organizations that have not yet published required disclosures face immediate compliance gaps that could result in regulatory action or customer complaints.

Managing switching requests for AI workloads

AI workloads present particular complexity for switching operations due to the interconnection between data, trained models, and operational configurations. Organizations must ensure that switching procedures preserve not just raw data but also the metadata, documentation, and configuration information necessary for customers to operate AI capabilities in new environments.

Documentation requirements under the EU AI Act intersect with Data Act switching rights. Article 53 documentation packages should accompany AI workload migrations so customers and regulators can trace model provenance, training data characteristics, and validation results after switching. Organizations must coordinate switching procedures with AI compliance teams to ensure documentation transfers appropriately.

Data format considerations for AI workloads include model weights, training datasets, evaluation datasets, configuration files, and operational logs. If you are affected, define standard export formats that preserve data utility and enable receiving platforms to load and operate AI capabilities without extensive transformation. Format documentation should explain data structures clearly enough for technical teams at receiving organizations to process imports correctly.

Fee transparency and Article 25 compliance

Article 25 mandates transparent fee structures for switching assistance services. Organizations must publish clear information about any charges associated with switching, including data export fees, technical assistance charges, and transition support costs. Fee schedules must be accessible to customers before they commit to services, enabling informed decisions about provider selection.

The Data Act sets up a trajectory toward eliminating switching fees entirely. Organizations may charge reasonable fees reflecting actual switching costs during transitional periods, but must reduce charges progressively. By 2027, switching fees must be eliminated. If you are affected, plan fee structures that comply with current requirements while preparing for the transition to zero-fee switching.

Fee transparency extends to indirect costs that might affect switching decisions. Organizations must disclose any charges that could apply during switching processes, including early termination penalties, data extraction fees, or technical support charges. Hidden costs that emerge during switching processes may violate transparency requirements and damage customer relationships.

Smart contract controls and Article 30 safeguards

Article 30 requires that smart contracts used in data sharing arrangements include termination controls—commonly called kill switches—that enable authorized parties to pause or stop automated data sharing when circumstances warrant. These safeguards ensure that automated processing does not continue inappropriately when business relationships change or security concerns arise.

Implementation of Article 30 safeguards requires technical controls that can halt smart contract execution cleanly without corrupting data or creating inconsistent states. Organizations using smart contracts for data sharing must verify that kill switch functionality works reliably across different scenarios and that recovery procedures can restore appropriate states after termination.

Audit logging for smart contract operations must capture sufficient detail to show compliance and support dispute resolution. Logs should record authorization decisions, data transfers, fee calculations, and any interventions that pause or end processing. These records provide evidence of proper operation and help resolve disagreements about smart contract behavior.

Day-one monitoring and operations

The first days and weeks after application date require intensive monitoring to identify issues before they escalate. If you are affected, establish real-time dashboards tracking switching request volumes, processing timelines, and any exceptions or escalations. Early identification of operational problems enables rapid response before regulatory attention or customer dissatisfaction intensifies.

Support team capacity should be elevated during the initial application period to handle increased inquiry volumes. Customers may have questions about their new rights, how to exercise switching options, or clarification about fee structures. Adequate support staffing prevents delays that could be perceived as obstruction of switching rights.

Vendor and partner readiness should be validated as customers may need to coordinate switching activities with multiple parties. If you are affected, confirm that third-party service providers, data processors, and technology partners can support switching procedures and meet required timelines. Vendor bottlenecks could delay switching completion and create compliance exposure.

Coordination with EU AI Act obligations

Organizations operating AI systems subject to both the Data Act and EU AI Act must coordinate compliance activities across both regulatory frameworks. Data Act switching rights interact with AI Act documentation, transparency, and governance requirements. Switching processes should preserve AI Act compliance status rather than creating gaps that require remediation.

Technical documentation under AI Act Article 53 should transfer with AI workloads during switching. Customers receiving AI capabilities need documentation to understand system characteristics, maintain compliance, and operate safely. If you are affected, establish procedures ensuring documentation completeness and accuracy during switching operations.

Ongoing monitoring requirements under both frameworks continue after switching. If you are affected, coordinate with customers to ensure monitoring handoffs preserve continuity and that any compliance obligations transferring to customers are clearly communicated. Post-switching support should include guidance on maintaining compliance in new operating environments.

Recommended actions for the next 30 days

• Establish monitoring dashboards tracking switching request volumes, processing timelines, and exception cases.
• Verify that fee schedules are published and accessible to customers as required by Article 25.
• Test smart contract kill switch functionality and document termination procedures for Article 30 compliance.
• Train customer service teams on Data Act rights and switching request handling procedures.
• Coordinate with vendors and partners to ensure switching support capabilities are operational.
• Review AI workload switching procedures to ensure AI Act documentation transfers appropriately.
• Elevate support capacity to handle increased inquiry volumes during the initial application period.
• Document lessons learned from early switching cases to refine procedures and identify improvement opportunities.

Key takeaways

The Data Act application date represents a fundamental shift in the relationship between data processing service providers and customers. Organizations must transition from viewing customer relationships as proprietary assets to providing genuinely portable services that customers can exit without undue burden.

Organizations that have prepared thoroughly for application day will find the transition manageable, while those that delayed compliance investments may face operational disruption, customer dissatisfaction, and regulatory scrutiny. The first weeks after application provide important learning opportunities that you should capture to improve procedures.

Recommended: treating the application period as an extended go-live event with appropriate operational support, monitoring, and rapid response capabilities. The investments in switching infrastructure and procedures will continue paying dividends as customer expectations for portability and flexibility increase across the data processing services market.