In this presentation, we will dive into a solution developed for supervisors at a municipality when they are about to conduct an employee assessment. We combine the power of Fabric and OpenAI to streamline and enhance the assessment process.

It is common for supervisors to spend a lot of time gathering quantitative information before the assessment and then analyzing and interpreting the data. We will explore a solution that uses the data processing power of Fabric and the analytical and interpretative capabilities of large language models to conduct a fair and accurate assessment and provide comprehensive and insightful feedback.

We will employ techniques like RAG (Retrieval Augmented Generation) to store guidelines for the supervisors, and we will use Notebooks and Python in Fabric with tools such as Semantic Link to retrieve data from semantic models (Power BI) and PySpark to retrieve data from Lakehouse. We will send the data to an LLM with an appropriate prompt to generate a document that guides us through the assessment.