Generative AI Custom Evaluation in SAP AI Core

The concept name was not provided, but based on the sources, this appears to relate to Custom Evaluation / LLM-as-a-Judge in SAP AI Core. This is a framework within SAP AI Core that allows developers to define and run custom evaluation metrics for generative AI outputs, including grounding and prompt optimization scenarios. Developers use it to assess the quality of large language model responses by configuring custom judges — including other LLMs acting as evaluators — against criteria such as tool-calling accuracy or retrieval-augmented grounding. It supports both quickstart workflows and comprehensive evaluation pipelines, enabling iterative improvement of prompts and model behavior within the SAP AI Core environment.