AI-Powered Classification of Job Postings for BUHL
- Custom Software Solution
- Automated AI Classification
- Optimization and Further Development

Order
The Augsburg-based BUHL Group specializes in industry services for the hotel and restaurant sectors. Its core competencies include staffing solutions, business services, media projects, and the operation of web portals. The company receives several hundred unstructured job postings in plain text format via various interfaces from partner companies. These postings are published on BUHL’s own portal.
Previously, the time-consuming process of classifying and structuring these ads was done manually by employees. Therefore, an AI-based solution was needed to automate this task and perform it quickly, efficiently, and reliably.
Challenge
A particular challenge in this project was that BUHL uses its own customer-specific parameters and rules for classification, which are subjective or industry-specific and therefore cannot be correctly implemented by even the most advanced AI without further adaptation. In addition, special attention was paid to data quality.
Solution
Implementation of a Custom Solution Using a RAG Framework
RAG (Retrieval-Augmented Generation) is an AI approach that combines text generation with targeted information retrieval. In our project with BUHL, RAG is used to retrieve customer-specific classification criteria and a number of customer-reviewed correct examples from a vectorized knowledge base and to apply them to the classification of new ads. This enables the system to make precise and context-appropriate decisions even with complex or new inputs. The continuously updated vector data ensures consistent results and enables adaptive refinement of the classification logic.
Optimization via a Testing Framework
A specialized testing framework was implemented to compare different versions of the solution and to improve the system in a systematic, sustainable, and documented manner. Various system prompts, instructions, AI models, and vector database models were efficiently tested using this approach.
Conclusion
The use of AI and RAG technology makes the classification of unstructured job postings efficient, accurate, and customizable. Customer-specific rules and examples in vector format ensure high data quality, reduce manual effort, and enable continuous optimization of the solution.

basecom's AI-based solution enables significantly more efficient classification of job postings. Our customers report high reliability and emphasize that the system can be continuously improved and optimized based on their ongoing feedback.

We have been working with basecom for over a year now and are very satisfied with both the quality of their services and the fact that we work together as equals. The basecom team has thoroughly impressed us with their expertise and well-thought-out concepts during the development of our online store. We look forward to continuing our collaboration.
Alexander Pholers, Online Project Manager, Pöppelmann GmbH & Co. KG Plastics Plant—Toolmaking Division
Company Information
The BUHL Group, headquartered in Augsburg, is the market leader in industry services for the hotel and restaurant sectors. Under the umbrella of BUHL Holding, BUHL Personal, STUDENTpartout, HOGAPAGE Media, and BUHL Services operate independently. With approximately 2,000 employees, the group offers a wide range of jobs and apprenticeship opportunities nationwide.

Jonas Dambacher, CSO
We look forward to meeting you.

Jonas Dambacher, CSO
