Evaluate results with AI chat (Results AI)

Open the evaluation view of the feedback process. There you find a dedicated Results AI tab. You only see it with HR access or management access, and the process needs completed text answers. With purely quantitative questions, there is nothing meaningful for the chat to work with.

The first time you open it, a consent page appears. It explains what the AI analysis does, that results are not deterministic and do not replace professional judgment, and that survey data is only transmitted to a technology partner under a contractual agreement (use for training purposes is contractually excluded). Confirm with your password or via SSO. After that, this step no longer appears.

At the top of the chat area, use the Analyze: dropdown to choose what is evaluated. The default is All text questions. Pick a single text question when you want to analyze it more deeply without other answers influencing the result.

Type your question into the input field and send it. During the analysis, the AI shows which step it is currently working on; then the answer appears in the chat history. You can ask as many questions in a row as you like. Questions in natural language give the best results, for example:

• What are the most common positive remarks about collaboration?
• Which development topics are mentioned most often?
• Summarize the critical points about team communication in three sentences.

The chatbot respects the active filter of the evaluation view: if a department is filtered, for example, the AI only analyzes those answers. And to keep things in perspective: AI answers are not a substitute for reviewing the original answers yourself. For important personnel decisions, check the answers directly in the evaluation.