Why AI Adoption Begins With People

Veronika Rubahn: When people talk about AI, the technology itself often comes first. What can a tool do? Which processes can be automated? Where can efficiency be created? These are important questions. But they are not enough. In the end, people decide whether a technology becomes part of their everyday work. Do I trust it? Do I understand its value? Do I feel that it supports my work? 
This is where the psychological perspective comes in. Technology acceptance does not only depend on whether a solution works well from a functional point of view. It is also shaped by expectations, experiences, social influences, framework conditions and perceived usefulness. These factors have been studied in psychology for a long time. What makes this exciting is applying these insights to AI and to specific business contexts.

Veronika: For me, it means not looking at AI as an isolated technology. You need to understand how people work, which tasks they have, where they see potential and which concerns they have. 
This is particularly crucial when it comes to AI, because acceptance is not just a technical question. It is also about how people understand information, how they make decisions, what value they perceive and whether a solution fits into real-world working processes.

Veronika: The starting point was the question of where AI can truly create value and how potential solutions can later be accepted. To explore this, I conducted interviews with employees from different functions and departments. I wanted to understand how they work today, where they already use AI, where they see potential and which tasks are especially time-consuming or challenging. Questions, uncertainties and concerns were just as important.
These conversations led to concrete use cases. Some had already been tested, while others were ideas for the future. At the same time, it became clear under which conditions these use cases could work. So it was not only about creating a list of possible applications, but also about better understanding the conditions required for AI adoption.

Veronika: What was interesting was that certain topics were repeated across different areas. Although the roles and tasks were different, similar areas of potential use came up several times. It was therefore worth looking beyond individual teams to identify patterns across the organisation. 
Another important learning was that AI is not automatically the best solution. The aim is not to develop an AI application for every problem. Sometimes another solution is simpler, safer or more effective. This became clear, for example, when dealing with sensitive data. If information must first be anonymised or rewritten with considerable effort before it can be entered into an AI tool, the effort may outweigh the benefit. Insights like these help companies use AI realistically and responsibly. 

Veronika: After the interviews, we prioritised the use cases in a workshop. One of the methods we used was an impact-effort matrix. This looks at the potential value of a use case and the effort required to implement it. 
In addition, we used a success criteria grid. This allowed us to evaluate the use cases based on specific criteria. For example: Is the necessary data available? Are there technical or organisational requirements? Is the use case realistic to implement? The aim was not to decide purely based on gut feeling. Especially with AI, there are often many ideas and a lot of momentum. Structured prioritisation helps identify the use cases that promise clear value and are also feasible.

Veronika: A very important one. Expectations and concerns show what people need in order to accept a technology. Some people ask themselves whether AI will make their work easier. Others want to understand how reliable the results are or which data they are allowed to use. Others are unsure which tasks they will continue to perform themselves and where AI should support them. 
These questions should not be addressed only at the end of a project. They belong at the beginning. If you understand them early, you can derive concrete measures from them. For example, clear communication, guidelines, training or adjustments to processes. Adoption does not happen automatically just because a tool is available. People need to recognise the value, build trust and have the opportunity to integrate AI meaningfully into their everyday work.

Veronika: When AI is integrated into customer experiences, it changes the relationship between customers and organizations. AI can make interactions faster, more accessible or more personal. But it can also create distance if people do not understand how decisions are made or whether they still have control. 
From a psychological perspective, it is therefore not only the outcome of an interaction that matters, but also how people experience that interaction. Did I feel understood? Was it transparent what was happening? Could I intervene? Is there a person I can contact if the AI cannot help? AI should therefore be designed and introduced in a way that strengthens trust. This can be especially important in relationships with customers.

Veronika: Companies should involve people early. Not only to collect ideas for use cases, but to understand needs, pain points and potential resistance. Focus is also essential. AI opens up many opportunities, but not every idea should be pursued immediately. It is important to realistically assess where AI creates real value, which requirements are still missing and which risks exist. 
For me, the central point is this: AI can only have a sustainable impact if it is meaningful, understandable and trustworthy for people. Successful AI adoption therefore does not begin with the tool. It begins with people.