Organization settings
Admins manage organization-wide settings on the My Organization page. The tabs covered here are:
The Compliance tab, with the Audit Assistant and the AI Safeguards Report, has its own article. See Compliance and the Audit Assistant.
To change your organization’s name or your number of seats, contact Sanitized AI.
General
- Organization details: your Organization ID. Your IT team uses it when they deploy the extension to managed computers. Nobody types it to sign in.
- Onboarding: click Restart onboarding to run the guided setup again. Your current guardrails and AI sites stay as they are until you finish it. When you finish, your new answers replace them.
The Reporting lines check is on the Users page, where you fix the gaps it finds. See Set up reporting lines.
Protection
This tab holds the settings that change what people see when they use AI.
Monitoring mode
Turn on Monitor the whole organization to stop warning or blocking anyone on any AI tool. Prompts send normally, and anything a guardrail flags is recorded as a monitored event. This covers the browser extension and the desktop apps. Turn it off to enforce again. See Monitor a tool before you enforce.
AI usage policy
If your organization has an AI usage policy, link to it from every warning, so people see the rules behind a flag.
- In Policy name, enter the name of your policy, for example Acme AI Usage Policy.
- In Policy link, enter the web address where people can read it. The
address must start with
https://orhttp://. - Click Save policy.
The warning then shows your policy’s name as a link under Why this was flagged. To remove it, clear both fields and save.
Real details on screen
When someone sanitizes a prompt with fake data, the extension can show their real details in the AI’s reply, on their screen only. The AI tool only ever receives the fake data. See See real details in AI replies.
Show real details in replies is on by default. Turn it off to stop this for everyone. People can also turn it off for themselves.
Alerts for weakened guardrails
With Email administrators on, every admin gets an email when someone switches off a guardrail, deletes a custom guardrail, loosens its action or lowers its sensitivity. Turning it off sends one last email to every admin. This card appears only if Sanitized AI has turned these alerts on for your organization.
Data & privacy
This tab holds the settings for what events record, what is counted, and who outside your admins can read it.
File names in events
When a guardrail flags a file, the event records the file’s name. A name can
give something away, such as a client’s name or a matter number. To keep only
the file type, turn on Mask file names. An event then shows •••.docx
instead of Tremblay v. Roy - Retainer.docx, on the
Events page and in exports. If the file type isn’t a
common one, the event shows only •••.
The setting covers every file event, including files that people send without a scan and files that a guardrail only monitors.
Masking applies to events recorded after you turn it on. Events already recorded keep the name they were saved with, and turning masking off doesn’t bring masked names back.
What people use AI for
Turn on Count prompts by task and topic to see what your organization uses AI for on the AI Usage page. It’s off until you turn it on. See What people use AI for.
Sharing with your partner
If a partner, such as your IT provider, manages Sanitized AI for your organization, the Sharing with partner card shows what they can see. They always see usage totals and events with the data masked. They never see prompt text, file names, or what was found.
- Show people’s names: off by default. While it’s off, your partner sees
each person as a code, such as
user-3f9a2c1b, the same code every time. - Send high-severity alerts: on by default. Your partner’s staff get an email when a high-severity event happens in your organization, so they can respond.
Only your admins can change these. Your partner can see what you chose. The card appears only if a partner manages your organization.
Data export API key
A key that lets another system, such as a data warehouse, pull your events. See Export events.
Extension settings
Browser extension links
This card lists the store links for Chrome / Brave extension, Microsoft Edge extension, and Firefox extension, with a copy button for each. Share them with your team, or let the welcome email do it for you. See Roll out the extension.
If your organization is managed by IT, this card is replaced by the Managed extension setup guide for your IT team. See Deploy the managed browser extension.
Extension Session Duration
Decide how many days people stay signed in to the extension, from 1 to 90. The default is 7. New sign-ins use the new length right away. People already signed in switch to it the next time their sign-in renews.
The extension renews a sign-in automatically while the browser is in use, so most people only sign in again if they don’t use their browser for longer than this period, or if you sign them out.
Sign-in & provisioning
This tab controls how people sign in and how accounts are created:
- Access policy: which sign-in methods are available. See Choose how people sign in.
- Company domain approval: the email domains your organization owns. You need an approved domain for automatic account creation and for Okta.
- Enterprise SSO: your Okta connection. See Configure Okta SSO.
- SCIM provisioning: automatic account management from your identity provider. See Configure SCIM provisioning.
The text on this tab is in English only for now.