The Youth Welfare Office in Landratsamt Augsburg, an IBM client, is tasked with the support and protection of children, adolescents and families across the district. Caseworkers in this group deal with legal and financial issues related to youth welfare as well as integration assistance and childcare costs. In order to determine the appropriate cost contribution for each case under their purview, these individuals conduct extensive research from a variety of different data sources, such as online legal references, documents submitted by citizens and other supplementary documentation, before they begin work on each case. While the current process ensures each case is thoroughly researched, it is also quite time-consuming and can, at times, be error prone. Furthermore, there is a shortage of specialized professionals who are skilled in conducting this kind of research.​

To resolve these challenges, Landratsamt Augsburg began collaborating with IBM^® Client Engineering and IBM Consulting^®. During a series of workshops, the team identified their most pressing pain points and outlined a vision for how they could help free up welfare caseworkers to focus on their cases, rather than the manual research process. The team at Landratsamt Augsburg knew they wanted to be more efficient but given the sensitive nature of this casework, they were understandably cautious about adopting new technologies that could potentially introduce new questions and challenges. Some members of the team were eager to explore generative AI (gen AI) as a possible solution, while others were more skeptical. To test out its potential, the team joined a three-day hackathon with IBM.

During the three-day event, 8–10 engineers came together to build a customized web-application that was integrated with IBM watsonx^™ Assistant and supported by IBM watsonx.ai^™ and IBM Watson^® Discovery. It was designed to query multiple disparate data sources set forth by the client and summarize findings for each case. The tests conducted during the event demonstrated an up to 91% improvement in the time required to compile case data. Before the pilot, it could take up to one hour to compile all the various data points that inform a case whereas during the pilot, the team observed it could take as few as five minutes to compile these data. Anecdotally, the team also observed a meaningful improvement in the quality and accuracy of the case data summaries generated by the new solution. The enthusiasm following the pilot was undeniable. The team is eager to keep exploring where they can improve efficiency and performance across the group by deploying gen AI.