Why are culture and mindset important in the AI context?

“How is AI changing my workplace? Which tasks will it take over in the future? Such questions can trigger uncertainty among employees. Companies should take these concerns seriously and address them openly. AI can take over routine tasks and thus create free space for more demanding activities. A good AI solution only delivers benefits once it is accepted in everyday work. Large companies rely on change programs for this; SMEs usually achieve acceptance more quickly through open team discussions, clear rules and visible successes. Employees need to know how to use AI results and when to take a closer look. Culture & mindset are not a one-off project, but a continuous process: closely aligned with the respective AI use case, with clearly defined goals and effects on tasks and roles.

Culture and mindset become particularly important when a company:

• is introducing its first AI applications and wants to build acceptance in the team,

• notices that employees are already using or avoiding AI tools unofficially,

• significantly changes tasks or roles through AI,

• wants to reduce uncertainty or skepticism about AI in the team,

• wants to establish an open error culture when trying out new AI applications, or

• wants to systematically integrate employees’ specialist and experiential knowledge into AI projects.

The earlier you factor culture and mindset into your thinking, the easier it is to embed AI in everyday work.

In 5 steps to a good way of dealing with AI in the company:

• Step 1 – Look at actual usage
  Record where AI is currently used, avoided or applied with uncertainty (including unofficial tools). Distinguish: is knowledge lacking, are rules unclear, or is the solution unsuitable? Low usage is just a signal, not yet an acceptance problem.
  

• Step 2 – Use employees’ knowledge and set a clear goal
  Formulate a measurable goal (e.g. “Within 6 weeks, 3 service employees use the AI assistant and provide documented feedback”). Gather requirements from everyday work: what does the AI need to be able to do in order to really help?
  

• Step 3 – Clarify the rules and choose a narrowly defined use case
  Clear guardrails create security: the company should define what AI may be used for and which data are approved for this. Just as important is transparency about what will change in terms of tasks and roles. Where there is a works council, it is ideally involved at an early stage. For all those involved to use AI safely and interpret its results, appropriate training is needed. This has also been required by the EU AI Act since 2025 (more on this in the Compass topic Talents and Skills). For the start, a narrowly defined task is suitable that can be tested within a few weeks, in an environment where mistakes are seen as learning opportunities and do not lead to blame.

• Step 4 – Define responsibilities
  Appoint a contact person, for example a manager or a tech-savvy person from the team. This person manages the schedule and feedback, reports to management and quickly clarifies questions about rules or responsibilities.
  

• Step 5 – Evaluate after 4 to 8 weeks and draw conclusions
  Check goal achievement and the overall process: is production faster but corrections more time-consuming? Document the results and decide: continue, adjust, limit or stop. This creates experiential knowledge for the next AI steps