AI Act 2026: What Really Changes for Companies, Providers, and Deployers

AI Act: From Theory to Concrete Obligations

August 2, 2026 marks an important milestone in the implementation of the European AI Act.

The regulation entered into force in 2024; some provisions have been in effect since 2025; and now new rules are becoming applicable to companies, developers, suppliers, and professional users of artificial intelligence systems.

But what really changes?

In this webinar, Andrea Bosio of Zero11 discusses this with Maria Chiara of Studio Torta, addressing the AI Act from the operational perspective of those who develop AI, integrate it into business systems, or use it within their own processes.

Provider or Deployer?

One of the first issues to address is understanding one’s role.

Those who develop or customize an AI system for a client may assume the role of provider.

Those who professionally use an AI solution, on the other hand, are generally deployers.

These are two distinct roles, but both come with specific obligations, responsibilities, and requirements.

Topics Covered

The webinar delves into the following:

• the phased implementation of the AI Act;
• the difference between providers and deployers;
• the classification of AI systems based on risk;
• responsibilities along the supply chain;
• data quality and governance;
• prevention of bias and discrimination;
• AI literacy and training;
• use of third-party models;
• black-box and generative systems;
• privacy and data protection;
• documentation and quality control systems;
• obligations for developers of custom AI solutions;
• modifications made by the deployer;
• controls and penalties provided for by the regulation.

From Compliance Burden to AI Governance

The AI Act is not just about regulatory compliance.

It compels companies to ask fundamental questions about how artificial intelligence systems are designed, procured, integrated, and used.

This means implementing processes for governance, risk management, data quality, and accountability starting from the design phase.

This approach can transform a regulatory constraint into an opportunity to build more reliable, transparent, and governable AI systems.

What About Penalties?

The regulation provides for significant penalties, which can reach up to 35 million euros or 7% of global revenue in the most serious cases.

All the more reason not to wait until the last minute.

Watch the webinar and find out what questions your company should start asking today to use AI in a compliant, responsible, and sustainable way.