AI evaluations: Practical content standards for real LLMs

Content designers are increasingly responsible for AI-generated content they didn’t write. But how can you measure “good” when the system changes output each time, and how do you improve it?

Drawing on UXCC’s AI evaluation workshop, this session gives attendees a practical introduction to building evaluation frameworks that work with real-world LLMs.

In this session, you’ll learn:

• A practical approach to building evaluation frameworks for LLM outputs that hold AI content to real UX content standards.
• How to assess AI-generated content qualitatively — judging clarity, tone, and usefulness rather than relying on gut feelings.
• The basics of building custom evaluation rubrics you can adapt for your own team or product.

What you’ll need

To participate, you’ll need access to a Google account (for Sheets) and a free chatbase.co account.