Security review

Security questionnaire answers

The questions a security review usually asks first, answered with what the workspace does today and a link to the page that shows it.

Downloads

The same answers as a spreadsheet to paste into your own questionnaire, and the security page as a PDF.

Answers as a spreadsheet (English)Answers as a spreadsheet (Japanese)Security overview (PDF)

For a legal review, see the data processing addendum template and how the product uses AI.

Access

Does the product support single sign-on?
Yes. Your own identity provider can be registered by SAML or OpenID Connect, roles and groups can be given from its claims, and single sign-on can be required for everyone, with emergency administrators kept for a broken provider. More
Can accounts be provisioned and removed from our directory?
Yes, by SCIM. Deactivating someone in your directory ends their sessions and stops their computers. More
How are sessions protected?
Sessions end when idle, on a sudden change of browser or network, and when a person signs out everywhere. Sensitive changes ask the person to have signed in recently. More

Data protection

Is data encrypted at rest?
Stored credentials, saved logins, identity provider settings, webhook secrets and pictures of the pages a coworker saw are encrypted before they are written. Every backup snapshot is sealed with AES-256-GCM before it leaves the workspace. More
Can staff or the AI see our passwords?
No. People type passwords into a private form the coworker never sees, password fields are hidden from what it reads back, and a saved password is never shown again, to its owner or to an administrator. More
Where is our data kept, and for how long?
In the workspace's own database, with each kind of data and its retention set out class by class on one page. More
Which other companies receive our data?
They are listed in the privacy notice, with what each receives. More

AI governance

Is our data used to train AI models?
No. The product does not use prompts, conversations, files or results to train models. More
How is the AI kept from acting on its own?
Boundary rules are checked before every action and fail closed, consequential actions wait for a person's approval and expire if nobody answers, and password fields and payment buttons are always handed to a person. More
What about instructions hidden in web pages or documents?
Text a coworker reads is treated as information, not as orders. Attempts to slip in instructions are detected and flagged, and risky actions still wait for a person. More

Logging

Are actions logged, and can the log be altered?
Every action a coworker took and every one the rules refused is in the audit trail, with secrets masked. Entries are only ever added and each is chained to the one before it, so a changed or removed entry shows. More
How long are logs kept, and can we export them?
The audit trail is kept for good by default, or for 30 to 3,650 days as an administrator chooses, and a legal hold keeps what it covers. It can be exported, and an evidence pack bundles it for an auditor. More

Network

Can a coworker's computer reach anything on the network?
Each computer works under its own network policy. Whatever the policy says, link-local and cloud metadata addresses are refused, and outgoing messages and files are checked for secrets and personal data. More

Continuity

How is data backed up, and are restores tested?
When backups are on, a sealed snapshot is taken every 10 minutes when something changed, and once a week the newest one is restored into a scratch database and checked against the live one. More
Where can we see availability?
A public status page shows the workspace's health and its history. More

Privacy

How are requests from data subjects handled?
A member can download their data and request erasure from their settings, and administrators track privacy requests with their due dates. More
Is a data processing addendum available?
Yes, a template is published for review. More

Vulnerabilities

How do we report a vulnerability?
Through the vulnerability disclosure policy, which gives the security contact. More