EU AI Act: what changes for manufacturers using AI?

The EU AI Act is the world’s first legal framework for artificial intelligence. Many companies assume the law only matters for tech companies building their own AI models. That is a misconception. The law applies to every organisation using AI, even if AI is only used to unlock knowledge, document processes, or help employees find what they need faster.

At Taggl, we do not see the EU AI Act as a burden, but as a practical framework for responsible AI use. This knowledge article explains what the law means, which categories matter for manufacturers, and how our platform is designed around those principles.

For readers who want to verify the law at the source: the European Commission explains the AI Act framework, and the official legal text is available on EUR-Lex as Regulation (EU) 2024/1689.

The law works with risk levels, not blanket bans

The AI Act uses a risk-based approach. Not all AI is treated the same. The weight of the obligations depends on what an AI system does and how much impact it has on people.

For most AI applications in manufacturing, including what Taggl does, limited risk is the relevant level. The obligations are concrete and manageable, provided you know what they involve.

The simplification measures also clarify that AI functionality in products is not automatically high-risk when it only assists users or optimises performance, and a failure does not create a health or safety risk. That distinction matters for manufacturers using AI as knowledge or process support, not as a safety component that intervenes autonomously.

Taggl does not build models, and that matters

The AI Act distinguishes between two roles. A provider develops an AI system and places it on the market. A deployer uses an existing AI system in a professional context.

Taggl does not build its own AI models. We use AI as a component inside our platform: to process speech, build knowledge bases from SOPs and messages, and help employees quickly find the right information. Your organisation is the controller for your data. We are the technical party enabling that, with the related agreements and responsibilities.

What the law asks in practice

For manufacturers, the practical impact comes down to four themes: AI literacy, transparency around AI-generated content, GDPR arrangements for voice processing, and human control over decisions.

AI literacy

Since 2 February 2025, organisations are legally required to ensure that employees working with AI systems understand what those systems do and where their limits are. The European Commission explains this obligation on its page about AI talent, skills and literacy. This is not a one-off training requirement, but an ongoing responsibility.

We support this by being transparent about how Taggl works: what the AI does, what it does not do, and how output is created. That makes it easier for manufacturers to meet this obligation.

Transparency around AI-generated content

When employees ask a question through the Taggl chat assistant, or read a knowledge base page assembled from SOPs and Taggl messages, it must be clear that they are working with AI-generated information. This is a legal requirement under Article 50 of the AI Act.

Our platform shows this by default: AI answers and AI-assembled knowledge base articles are always clearly labelled. No hidden automation. Users can always see the source.

The new Parliament text also names 2 December 2026 as the date for machine-readable labelling or watermarking of AI-generated content. That is separate from the practical user-facing labels we already show, but it points in the same direction: AI output should be recognisable and traceable.

Voice processing falls under the GDPR

Taggl processes voice recordings. Speech is personal data, which means the GDPR applies. That is a different responsibility structure from the AI Act: the legal basis for processing sits with the employer as controller, documented in the employment relationship or internal policy. With Taggl, you enter into a data processing agreement that covers the GDPR framework.

Voice data does not fall under the high-risk categories of the AI Act in this use case. Taggl does not make decisions about employees based on voice analysis. The platform uses speech solely as input for knowledge capture.

No autonomous decisions about people

The AI Act places the heaviest obligations on AI that makes decisions with direct impact on people, such as access to work, finance, or healthcare. Taggl makes none of those decisions. The platform supports knowledge sharing and process assurance. Employees find what they need faster; the decision remains human.

The timeline at a glance

This does not mean the whole AI Act moves back: the European Parliament still names 2 August 2026 as the start date for most provisions. For AI systems used as safety components in regulated products, the agreement names 2 August 2028 as the new date. Manufacturers should therefore not only look at the legal deadline, but already clarify which AI systems are in use, which risk level applies, and where human control remains necessary.

What Taggl is built on

The AI Act asks for transparency, human oversight, and clear responsibilities. These are not new requirements for us. They are the principles our platform is built on.

• Visible labels. AI-generated content is clearly labelled as such, both in chat and in the knowledge base.
• GDPR framework. Voice processing is arranged through a data processing agreement and appropriate technical measures.
• Human control. The platform does not make autonomous decisions about employees or processes.
• Explainable use. Customers are informed about how the AI works and what its limitations are.

Want to see how Taggl preserves knowledge in your manufacturing environment and how we handle GDPR and the EU AI Act? Get in touch or request a demo.