In this session of Ziflow Academy, we'll show you how to harden your ReviewAI checks over time so they get more accurate the more you use them. You'll learn how to read Check Reliability Levels, use reviewer feedback to update a prompt to a new version, validate the fix, and understand how versioning applies to existing and active proofs.
Topics covered:
- Understanding Check Reliability Levels – What Perfect, High, Medium, Low, and Not Evaluated mean, and how they're calculated
- Updating the Prompt – Using feedback and common failure reasons to edit a check and create a new version
- Validating the Fix – Re-testing previously failing proofs to confirm the improvement
- Versioning & Retroactive Changes – Why prompt updates are forward-facing and don't alter closed proofs
- Resetting Checklists on Active Proofs – Manually applying the newest version to an in-progress proof
After this video, you'll be able to:
- Interpret a check's Reliability Level and know when a prompt needs work
- Use reviewer feedback to refine a prompt and save it as a new version
- Validate prompt changes by re-testing proofs that previously failed
- Understand how versioning affects closed and decisioned proofs
- Reset a Checklist on an active proof to apply your latest, hardened version
For more information, check out these resources in our Help Center:
- About ReviewAI
- Set up ReviewAI
- ReviewAI prompt description guidelines
- Create and edit ReviewAI prompt descriptions
- Define how ReviewAI writes suggested comments
- Run ReviewAI
- How ReviewAI uses your data
This video references features that are only available in certain Ziflow plans. For a clear breakdown of what’s included in your plan, visit ziflow.com/pricing.
Related to
Comments
0 comments
Please sign in to leave a comment.