Responsible AI policy
Published August 20, 2026
What the policy covers
This policy sets what our models may and may not do with your documents, and how a person stays in charge of every change.
Suggestions only
Models suggest edits, and a person accepts each one
Reasons shown
Every suggestion cites the playbook position or rule behind it
No training
Customer documents are never used to train a model
Evaluation
Each release is tested against contracts reviewed by lawyers
Signing
Proviso never signs or sends a contract on its own
Request the full policy
We send the full policy and our latest evaluation summary, usually within one business day.
