Putting AI to work for faster, safer vehicle inspections

In Germany, there are only few organizations entrusted with conducting millions of legally mandated technical inspections. Consequently, the market for licensing partners is highly competitive. GTÜ Gesellschaft für Technische Überwachung mbH (GTÜ) is the largest officially recognized monitoring organization for freelance automotive experts and inspection engineers. GTÜ partners conduct more than 4.7 million main inspections and over 8.3 million vehicle inspections annually—approximately every sixth vehicle in Germany. This makes GTÜ an important pillar of German traffic safety and environmental protection.

Demographic change presents the organization with a dual challenge in both the short and long term. The shortage of skilled workers, especially with nearly 20 million Baby Boomers retiring by 2036, requires that knowledge built up over the years be digitized and made accessible long term. At the same time, there’s a need to bridge the growing gap between older inspectors who adhere to older but familiar systems and younger digital natives who expect intuitive tools.

Throughout this process, partner interests remain at the forefront. Supporting the inspection engineers who work for GTÜ is the company’s top priority. The company’s 20-person support team receives over 60,000 calls annually.  An AI assistant was intended to relieve support staff of simple inquiries and enable them to focus on complex requests. Inspection engineers would also be able to complete their inspections faster, without waiting in a telephone hotline queue.

“We want to not only bring advantages to our partners but also excite them and show that we are passionate about what we do,” explains Lukas Tomalla, development manager at GTÜ.

Based on AWS Bedrock, GTÜ’s CoExpert is a holistic solution with four core components: an intuitive interface for user inquiries, a system that processes fed documents for the vector database, a feedback mechanism that ensures the quality of answers, and an observability module that provides insights into given answers to continuously make improvements. 

A crucial success factor was the use of Slalom’s proprietary AI accelerator as a framework, which enabled rapid implementation. 

From the beginning, knowledge transfer was a central concern. Slalom’s task was not only to deliver a solution but to enable GTÜ to implement and maintain AI applications themselves in the future. In the third project phase, Slalom transferred its knowledge to GTÜ employees who will maintain CoExpert and enable GTÜ to expand its AI applications.

The initial successes of CoExpert can already be observed. “In April 2025, for the first time in GTÜ’s history, we managed to stop the continuous increase in calls to the technical hotline,” Schaffors is pleased to report.

“Partnering with Slalom was an investment in our future,” says Thomas Emmert, CEO of GTÜ. ”With their data and AI expertise, they helped us to use existing data to create business value, build up long-term expertise, and take our employees with us on this journey. This is important for the long-term economic success of GTÜ.”

After the successful first project, GTÜ is already working on further innovations. The goal: just as in 1994, when GTÜ pioneered the first digital support system for inspection engineers, the company wants to position itself once again as an innovation leader in the German market with tools like CoExpert.