Tell us about your organisation so we can tailor the review process.
This helps us apply the right policies and processes to your reviews.
An API key is saved in this browser.
Enter your Anthropic API key — ask your admin if you don't have one. It's saved in this browser only.
Add standing instructions that are included in every review. Use this to encode local union policy, calibration preferences, or anything Claude should always know when reviewing your RAs.
Restore all preferences to their defaults. Your API key and review history are not affected.
Wipe everything — API key, org selection, custom instructions, history, and all preferences. The tool will restart as if opened for the first time.
Review Mode: Submit an RA and the tool will review and provide feedback on your RA.
Plan Mode: Start with an outline of an event. The tool will create a hazard summary, perfect for starting your RA.
Compare Mode: Compare two RAs, before and after feedback, to see whether students have taken on board your feedback.
How does it work? This tool uses AI to review risk assessments against standard union safety requirements and policy. It is designed as a first-pass review aid and should always be used alongside your own judgement.
Important: This tool uses API requests - each request costs real money. Please do not submit more requests than necessary. To view what the output of this tool looks like, without using tokens, please switch to demo mode in the menu bar.
Deep uses Claude Opus, which costs significantly more per review. Enter the coordinator password to unlock it for this session.