Intelligent and connected services have become essential in the manufacturing industry. The surge of these services has even started a new phase of industry; industry 4.0. The COALA consortium aims to develop a service to assist operators within this new industry. The COALA consortium is an European Union Programme and aims to develop a Digital Intelligent Assistant (DIA). The DIA supports operators in situations characterised by cognitive load, time pressure, and little or zero tolerance for quality issues with trustworthy AI components. Diversey is one of the consortium’s partners and perceives COALA’s DIA as a viable solution to their stoppage challenge on their production lines in the factory. These stoppages can be caused by various bottlenecks, which are hard to determine due to the production lines’ complexity and processes. Therefore, this thesis explores how to prepare for AI service adoption within the factory of Diversey. However, management from the Diversey factory has attempted to identify the bottlenecks by collecting production information through operator data entries. Unfortunately, the operators did not provide the tool with quality data entries. Therefore, replacing this tool with the DIA most likely not succeed, especially when the new tool utilises AI technology that requires learning data to generate insights. In order to explore this resistance toward new tools within the factory, I used the frame creation method of Dorst. This method excels at finding innovative solutions for problem definitions with previously unsuccessful attempts. The method explores underlying themes within the context to reframe the problem definition and find new solution spaces. The themes are formulated from interviews with management, eight support staff employees across four departments, a team lead, and six operators. Four themes are generated by analysing the values, interactions, and ‘currency’ exchanged between stakeholders. These themes regarding new tools describe the unclear contributions to operators’ work, the lack of acknowledgement regarding operators’ role and expertise towards production improvements. These themes result in a lack of trust between operators and management, which diminish the willingness to adopt new tools. Additionally, management expresses concerns about the themes expertise and consistency, which are essential to the manufacturing industry. In order to address these themes, COALA’s DIA is reframed as a stage for expertise and exploration. Instead of simply requesting data entries, operators are put in the spotlight where they can showcase their knowledge and expertise. Additionally, this stage provides a space for operators to explore production improvements through collaboration with COALA’s DIA. The framing of the problem definition addresses the data collection aspect of COALA’s DIA. However, AI systems change over time as the systems adapt to the data input. Additionally, the thesis did not assess the users’ perception of the current DIA interface. Therefore, I recommend further research into human-AI interaction with regards to the system’s evolution over time and the interface. Lastly, the themes are generated from one production line’s operators who are experiencing the most stoppages. The themes cannot directly be generalised to other production lines as the number of stoppages, or other contextual factors can influence the themes.