Book ↗ DE EN
Service: AI and the AI Act

AI and the AI Act, implemented soberly.

I help you implement the AI Act in your business: inventory, measures, training, technical implementation. Most SMEs only use AI and are therefore deployers. For them the effort is manageable once it is clear what applies and what does not.

Short answer

The AI Act is the EU regulation on artificial intelligence, Regulation (EU) 2024/1689. For SMEs, what matters most right now is AI literacy under Article 4 and the transparency obligations under Article 50. The latter have applied since 2 August 2026. Bracharz Consulting in Wilhelmsburg, Lower Austria, puts this into practice with an audit, a workshop, chatbots and content processes. Rainer Bracharz is a management consultant, not a lawyer.

What applies, what to do, what you can leave out

As of September 2026

Article 5 · since 2 February 2025

Prohibited AI practices

Certain AI practices have been banned in the EU since 2 February 2025. In the audit we check whether any of your tools falls under them.

Article 4

AI literacy in your team

Providers and deployers must support their staff in building AI literacy. The Digital Omnibus softened this duty in 2026: it used to say “ensure”, now it says “support”. A training session with certificates of attendance and an internal AI policy make this demonstrable.

Article 50 · since 2 August 2026

Transparency obligations

These obligations have applied since 2 August 2026. They were not postponed.

  • →Chatbots: people talking to an AI must be able to recognise it. This is the provider's responsibility.
  • →Generative AI: providers mark outputs in a machine-readable way. For systems already on the market before, this applies from 2 December 2026.
  • →Deepfakes: realistic AI images, videos and audio of real people, places or events must be disclosed by the deployer.
  • →AI text: if AI-generated text informs the public on matters of public interest, it must be labelled. Exception: a human has reviewed it, and a person or company bears editorial responsibility.

Annex III and Annex I

High-risk AI

For high-risk AI under Annex III the obligations apply from 2 December 2027. This covers, for example, AI in recruiting, creditworthiness or education. For AI in regulated products under Annex I they apply from 2 August 2028.

Your role

Provider or deployer

Most SMEs are deployers: they use AI but do not develop it. Anyone offering a chatbot under their own name or building one for clients may be a provider. Your role determines which obligations apply to you.

Leave it out

What you do not need

You do not need a large project. If you use AI and do not deploy high-risk AI, you mainly need an overview, clear roles, training and a policy. Fear does not help either. Breaches of the transparency obligations can cost up to 15 million euros or 3% of worldwide annual turnover, for SMEs the lower amount applies.

Source: AI Regulation (EU) 2024/1689, Articles 4, 5 and 50 and Annexes I and III, as amended by the 2026 Digital Omnibus. As of September 2026. This overview is not legal advice.

What you can commission

01

AI Act check for SMEs

For

Management of SMEs using ChatGPT, Copilot, chatbots or AI in marketing.

Content

I list every AI tool in your business and clarify your role as provider or deployer. Then I assess the risk class and show open transparency obligations.

Result

One document with the tool list, roles, prioritised measures and owners.

02

AI literacy for your team (workshop)

For

Teams that use AI in their daily work.

Content

What AI can and cannot do. Handling customer data under the GDPR. The labelling obligations. An internal AI policy.

Result

A workshop with certificates of attendance and a draft AI policy. This documents that your business supports AI literacy under Article 4.

03

AI chatbots with built-in transparency

What

Chat and search assistants for websites, for example for product search, appointment requests or first consultations.

Included by default

The AI notice under Article 50 is always included. The bot introduces itself as an AI. There is a handover to a human, and data protection is documented.

04

Content engine with a clear approval process

What

AI-assisted production of articles, landing pages and social media content, optimised for Google and AI search (SEO and GEO).

AI Act built in

Every text is reviewed and approved by a human. Editorial responsibility lies with you or with Bracharz Consulting. AI images and videos showing real people or places are labelled. Every approval is logged.

Benefit

More content for the same effort, without labelling risk and without interchangeable AI filler.

How it works

01

First call

30 minutes, free of charge. You tell me which AI you use or plan to use. I tell you honestly which parts the AI Act affects.

02

Audit and proposal

You receive a prioritised list of measures. It comes with a proposal with a clear scope and timeline.

03

Implementation

Training, policy, labelling or technical implementation. Direct, documented and without an endless project.

Three things in one pair of hands

I am a management consultant, not a lawyer, and I do not provide legal advice.

Related

Data Protection & Compliance

GDPR and the AI Act work together, for example when customer data goes into AI tools.

GEO: Get recommended by AI

The content engine in action: content that is found on Google and in AI answers.

Let's talk Business.

The first call is free. Afterwards you will know whether it fits.