AI
Artificial intelligence as Accelerators
AI makes recruiting faster and more accurate. However, it does not replace judgment or responsibility. The decision remains with the human being – and we help you to align the EU AI Act with your own business.
As we ourselves with AI is working
Recruitment is a sensitive area under the AI Act. Therefore, we clearly separate who is responsible for what.
Speed and preselection
- Matching and screening as support, not as a substitute
- Faster pre-selection and better hit rate
- Help with job profiles, interviewing, and evaluation
- In compliance with Swiss GDPR and GDPR
Judgment and Responsibility
- Every recommendation is personally reviewed by us.
- Assessment of motivation, fit and team
- We conduct the interviews with the candidates ourselves
- No automatic cancellation without human control
EU AI Act
Not whether you use AI – but for what
Since August 2, 2026, Regulation (EU) 2024/1689 is generally applicable. The AI Act is not a ban on AI, but rather a framework. It distinguishes between risk, and this risk arises from the specific application case, not from the product name.
An AI that formats a job posting is different from an AI that sorts 200 applications based on suitability. The same tool, two completely different risk levels.
Even those who only use a ready-made tool may have obligations. The AI Act distinguishes between providers and operators – you do not have to develop AI to be affected.
Four Risk levels
The AI Act does not treat every application equally. This classification is the starting point for any categorization.
Prohibited practices
It is not the technology that is prohibited, but certain applications – such as manipulative systems, exploiting vulnerabilities, or social scoring. This applies since February 2, 2025.
High-risk applications
Allowed, but with strict requirements: risk management, data governance, documentation, logging, human oversight, accuracy, and cybersecurity. This applies, among other things, to employment and personnel management.
Transparency obligations
People need to know when they are interacting with AI or when content has been generated or modified by AI. This applies to chat bots as well as generated text and images. Effective since August 2, 2026.
Minimal risk
The vast majority of everyday life. There are hardly any AI Act obligations here – however, data protection, copyright, employment law, and confidentiality apply regardless of this.
Does that also apply to Swiss companies?
It is possible. What matters is not just the company’s headquarters, but where AI is offered and deployed – and where its results are used.
A purely Swiss initiative
A company that uses AI exclusively for its activities in Switzerland does not fall under the AI Act solely because of this. Swiss data protection, employment and contract law continue to apply regardless of this.
With EU connection
- They offer an AI system to customers in the EU
- The results of your AI system will be used in the EU
- They have a branch in the EU that uses AI
- They bring an AI system to the EU market
Conversely, however, the following also applies: Not everything that has something to do with the EU falls under the AI Act. Article 2 lists specific points of reference. Therefore, when dealing with Swiss clients, we first examine the geographical context – and only then assess the risk classification.
The Schedule
As of September 2026. Transitional rules are being developed and should be reviewed again before any decision is made.
Our AI Act Check
Six steps, individually or as a package. At the end, you know which AI you can use, what you need to consider, and what your employees need.
AI inventory
We record which tools are actually in use, in which department, for what purpose, and with which data. The result is an AI registry.
Risk classification
We assess your role – as a provider or operator – and assign each use case a risk level. Including EU-related aspects.
Action plan
A clear list instead of an expert opinion: What should be done immediately, what can wait, and what is not necessary.
AI Directive
Simple internal rules that fit your business: approved tools, handling of data, approvals, reporting channels. No 50-page document.
Employee training
AI competence according to Article 4 – based on your actual work cases rather than on a hundred slides. What can be entered, when a result must be reviewed.
Regular check
Tools change, new ones are added. We keep the overview up to date, instead of creating it once and then forgetting about it.
This overview is an information and working basis and does not replace individual legal advice. Whether a specific AI system falls under a particular rule depends on the actual system, its function, and the current legal situation. For high-risk applications, data protection issues, or applications that are sensitive from a labor law perspective, we recommend that you additionally seek qualified legal advice.
Contact
Ready for yours First conversation?
In just about 30 minutes, we get an impression of your open vacancies, wishes, and expectations of us as a talent acquisition partner; or, if you are looking for a job, we can take your expectations into account for your next career leap.
Of course, we will also determine how we can best support you in this process.
Mo–Fr, 8:00–18:00
Write via WhatsApp Company address: creaSKILL GmbH, Natternweg 3, CH-4852 Rothrist
Correspondence address:
creaSKILL GmbH, Kaiser-Joseph-Str. 254, D-79098 Freiburg
Talent Solution Praxmarer, Rohrbachstraße 49a, AT-6060 Hall in Tirol