Built for client confidentiality.
Private by default, never used to train AI models, and walled off from anyone who shouldn’t see it.
Four promises
In place at launch, and in your agreement.
Confidential by default
- Chats private to each person
- Never used to train models
- Export or delete it, any time
Walls between matters
- Set by matter and by person
- Enforced across casewise
- Every change logged
Only the right people
- Only the people you add
- Chats stay with each person
- Admins can’t read them
Every action recorded
- Logged with who and when
- Find by person or matter
- Export it for your own review
Independent standards
Each gets its place here as it is awarded.
See it in the app. Your data, in Settings: never used for training, where it’s stored, export and delete.
Try it in casewiseEvery commitment
The detail behind the promises.
- Privacy
- Chats private to each person; matters shared only with their team; no admin access to chats.
- Model training
- Your material never trains the models behind casewise; providers process prompts under contract with no training rights.
- Encryption
- TLS in transit, and encryption at rest across databases and file storage.
- Data isolation
- Each account is logically separated, and nothing is reused across accounts.
- Access control
- Least privilege into production, plus roles and permissions inside group accounts.
- Secure development
- Peer review and automated scanning on every change, with regular dependency audits.
- Retention and deletion
- How long material is kept, and the deletion window, both stated in the agreement.
- Incident response
- A written plan with escalation paths, and notice to affected accounts on a confirmed breach.
- Hosting and residency
- Where material is stored and processed, confirmed in writing on request.
- Model providers
- Named providers process prompts under contract, with no training rights over your material.
- Backups and recovery
- Automated backups, with monitoring and alerting on unusual activity.
- Sub-processors
- The full list of third parties and what each one receives, published and kept current.
- Independent assurance
- Any formal audit or standard, stated only once held, with documents on request.
- Ethical walls
- Set by matter and by person, enforced everywhere in casewise, every change logged.
FAQ
Is my work used to train models?
No. We don’t use your prompts, uploads or outputs to train casewise or other AI models, and the providers we use work under contract with no training rights.
Who can see my chats?
Your chats are private to you unless you share them, and no admin at your firm can read one. We review conversations to find and fix problems. Your prompts and documents go to the AI providers named in the Privacy Policy, and some of them are overseas.
Is my clients’ information kept confidential?
Your chats, files and matters are private to you unless you share them, walls screen people out of a matter, and nothing you add is used to train models. Where your material is stored is confirmed in your agreement, on request.
Is my work encrypted?
Yes. TLS in transit, and encryption at rest across databases and file storage, with least-privilege access into production.
What is an ethical wall in casewise?
A screen between a person and a matter. Set it by matter and by person, and the matter disappears for them from their lists, search, the Vault and casewise’s answers, with every change logged.
Can I delete everything?
Yes. Delete a chat from its own menu, or delete your whole account from Settings, Account & security.
Doing a security review?
Send us your questionnaire and we’ll answer it in writing, item by item.