A project with the company Kollmorgen AMS, where we investigated how to design AI recommendations in automated guided vehicle (AGV) systems. The project focused on balancing user trust, motivation, and over-reliance and resulted in 17 design guidelines and a Figma prototype.
As AI becomes increasingly integrated into various systems, it brings new aspects to consider in Human-computer interaction (HCI). This project was in collaboration with Kollmorgen AMS, a global leader in navigation and fleet control of AGVs. It is important for them to keep up with new technology like integrating AI. This integration brings new user experience challenges, requiring new aspects to be considered. Kollmorgen AMS therefore needed to gain understanding for the expert users perception and response to AI recommendations in AGV systems to successfully implement it.
The project focused on how the introduction of AI recommendations into the UI of AGV guiding software affects the user experience, as well as how these effects can be reformulated into guidelines. To understand the problems and propose solutions the following research questions were acknowledged:
What factors are most important to provide expert users with a good user experience when implementing AI recommendations in AGV systems?
What guidelines should be followed when designing an interface balancing trust and facilitating decision making when using AI recommendations?
The project began with an initial pre-study, focusing on gathering a solid foundation based on contextual information retrieval and literature search.
Transparency, control, and clarifying the system is important to enhance trust and facilitate decision-making.
"I want to control when to get input"
"It helps a lot to have experienced colleagues"
"I learn on my own through trial and error"
Seven interviews were conducted with current or former application engineers to discover aspects about: Their work role, Working with layouts, Thoughts of new features, and Attitude towards AI.
In addition to interviews, observations were conducted to gain knowledge about the working steps, the participants mood and reactions during the process, and any problems that arise.
Realizes that they could have done it more efficiently and saved time, but says they
forgot about it.
In addition to interviews, observations were conducted to gain knowledge about the working steps, the participants mood and reactions during the process, and any problems that arise.
Realizes that they could have done it more efficiently and saved time, but says they
forgot about it.
