Many companies already use AI today — often as a matter of course. Sometimes as an assistant in daily work, sometimes as a tool in marketing, sales, or customer service. What has been possible for a long time will change in the coming years.
With the EU AI Act, it is regulated for the first time how AI must be organized, documented, and responsibly deployed. From 2026 at the latest, it will no longer be enough to use AI only “on the side.”
This article provides a compact overview:
- What the AI Act means in brief
- Which topics become important for companies
- What this specifically means for SMEs
- And what role an AI communication assistant like LIVOI plays in this
What Is the EU AI Act in Brief?
The EU AI Act is the European AI regulation. Its basic principle is simple:
The higher the risk of an AI deployment for people and fundamental rights, the stricter the requirements.
Not every AI automatically qualifies as a high-risk system. But: All companies that seriously use AI must address responsibility, transparency, traceability, and data handling.
Timeline (Simplified)
- Since 2024: The AI Act is in force.
- From 2025: First obligations take effect, such as AI literacy and prohibited practices.
- From August 2, 2026: The AI Act becomes applicable in large parts — especially for systems that interact with people or influence processes.
For SMEs, this means: The next two years are the phase in which structures and rules should be established.
Which Topics Become Important for Companies?
In practice, the requirements can be bundled into three core areas.
1. AI Literacy and Clear Responsibilities
Employees should be able to understand and responsibly use AI. This requires:
- Training and basic understanding (AI literacy)
- Simple guidelines for usage
- Clear responsibilities
At the same time, it must be defined which tools are permitted, which data may be used, and who is responsible for AI topics.
2. Transparent and Traceable AI Usage
Wherever AI interacts with customers, employees, or partners, it should be clear:
- that AI is being used
- what task it performs
- at what point a human takes over
Companies should generally be able to explain why a response or recommendation was generated — through protocols, documentation, and audit capabilities.
3. Deliberate Selection of Use Cases
Instead of “AI everywhere,” the focus is on targeted, well-controllable areas of deployment. Typical entry points are:
- Initial contact and qualification
- Appointment scheduling
- Recurring inquiries (FAQ)
The key is to start small, define clear handovers between human and AI, and make effects measurable.
What Does This Concretely Mean for SMEs?
The AI Act is not an AI ban. It is a step toward professionally operated AI systems.
For mid-market companies, this means:
- AI becomes part of organizational reality, not just an experiment.
- It is no longer sufficient for individual employees to use various AI tools “on the side.”
- Those who create structures early can deploy AI securely, traceably, and scalably — instead of having to react under pressure later.
What Does This Have to Do with LIVOI?
LIVOI is a company-specific AI communication assistant. It is deployed where many conversations, inquiries, and coordination tasks take place — for example:
- In sales
- In customer service
- In onboarding
- For internal inquiries
In the context of the AI Act, three points are particularly relevant.
Transparency
Companies clearly define:
- Where LIVOI provides support
- How communication takes place
- When a human takes over
Control and Data Sovereignty
LIVOI works exclusively with the knowledge of the respective company — not with arbitrary internet data. The data:
- Remains in a protected environment
- Is separated and encrypted per customer
- Is not used for third-party training
Traceability
Conversation histories can be documented, next steps derived, and processes cleanly integrated. LIVOI is not a black-box system but a building block in clearly defined workflows.
In short: LIVOI fits well with the direction the AI Act prescribes — AI that fits into processes, rather than AI that only impresses as a demo.
Three Steps You Can Take Now
- Take stock: Where are you already using AI today — consciously or unconsciously (Office features, CRM, chatbots, external services)?
- Define rules: Which tools are permitted, where may data go, who bears responsibility?
- Choose a clear use case: For example, initial contact, appointment booking, inquiry qualification, or internal FAQ support.