About the KI LAB Heidelberg:
The KI Lab Heidelberg supports companies in efficiently integrating Artificial Intelligence (AI) into their processes. It offers quick, cost-free implementation of AI pilot projects in collaboration with AI startups. Companies are guided through a five-week process that includes a potential analysis, scouting for startups, and the joint development of an AI pilot. This initiative, funded by the Ministry of Economic Affairs, Labour, and Tourism of Baden-Württemberg, aims to make AI applications more accessible to businesses.
Learn more about the KI LAB Heidelberg here: https://www.ki-lab-heidelberg.de/
As part of the portfolio of the Heidelberg Technology Park, the KI LAB Heidelberg plays a vital role in breaking down barriers between AI and practical applications, helping businesses experience the added value of AI in real-world scenarios.
Paul Becker, the head of the KI LAB Heidelberg, asked his clients:
"Time-consuming, expensive, and too complex – these are the most common reasons companies hesitate to adopt Artificial Intelligence."
However, based on experience from over 25 pilot projects between startups and established companies, we know this: AI can be deployed in businesses within just five weeks.
A Success Story: Junker Filter GmbH
It’s July 2023. ChatGPT is the hot topic in the tech world. Handelsblatt and journalist Larissa Holzki publish an 80-page special issue on AI. The AI hype is in full swing.
At that time, only a few had grasped the potential added value of AI and large language models for businesses. Many were still sitting in front of their computers experimenting with ChatGPT.
It was during this period that I met Julian Nuss, Managing Director of Junker Filter GmbH, at an event hosted by the KI LAB Heidelberg.
We struck up a conversation:
"I’ve been grappling with knowledge management for a long time. How can I document, organize, and make internal knowledge intuitively accessible to all employees?"
Since the advent of ChatGPT, Mr. Nuss had realized that large language models (LLMs) could provide a partial solution to this question. Companies like OpenAI and Aleph Alpha provide the technology. Together with Frank Rust and Andreas Schmidt from Neunzehn Innovations, the development began.
It’s worth noting: AI applications based on LLMs are still uncharted territory for small and medium-sized enterprises (SMEs). The approach was to simply start, experiment, and iterate – knowing that some ideas would end up in the trash.
The Result:
Within a few months, the first prototype of the knowledge agent "JUFI" was developed. This prototype has since been refined and integrated into daily operations. JUFI is a chatbot that provides answers to company- and industry-specific questions based on information and files stored in the company’s internal databases.
What impressed me the most?
At Junker Filter, it’s all about action, not just talk! Whether you speak to the executives or walk through the production halls, you see a hands-on approach everywhere. The company has become a front-runner in adopting large language models.
Back in July 2023, I wondered how many companies would recognize the trend and implement LLM-based solutions.
The exciting part? The KI Lab Heidelberg is currently supporting five more SMEs with LLM projects – more than I had expected back in July 2023. And the numbers keep growing.
But see for yourself:
The reference project by Junker Filter GmbH.
