The Virtual Service Technician provides engineers with direct access to relevant information through an interactive chat function. When a technician is on-site at an installation experiencing a fault, they can submit their question to the tool, which acts as a link between the machine, all associated data, and the technician as the end user. The tool draws its knowledge from real-time sensor data as well as company documents such as manuals, fault reports, and maintenance logs. Work is also underway to incorporate the experiential knowledge of seasoned technicians, for example through video blogs and technical drawings.

AI expert Fausto Albers of the HvA explains that the tool was developed specifically to bridge the knowledge gap between experienced and less experienced engineers. Within the SME and technical manufacturing sector, a great deal of valuable knowledge resides with the people who work there. Through new forms of AI, that knowledge can be captured and made accessible to others, enabling younger or less experienced technicians to work independently more quickly.

Refrigeration installation company IJskoud is one of the practical partners in the research. Director José Komin notes that refrigeration systems are becoming increasingly complex while skilled professionals are in ever shorter supply. IJskoud already worked with data and monitoring to prevent breakdowns, and the Virtual Service Technician fits closely with that approach. The tool monitors installations around the clock, uses data to predict problems, and provides more targeted support to technicians when they arrive on-site.

The collaboration between the HvA and IJskoud is mutually reinforcing. Dozens of students contribute to the tool, and the data collected by IJskoud is validated within the research framework, which increases its reliability. Albers emphasises that feedback from technicians in practice is indispensable: only through actual use can it be assessed whether the system’s outputs are correct, which is necessary for the system to become self-learning. Komin adds that better calibration of installations leads to greater efficiency and a longer operational lifespan, which benefits companies, people, and the environment alike.