Skip to content
Sanitized AI Help Center

Admin setup checklist

Sanitized AI sets up your organization and your admin account. You get a welcome email titled Welcome to Sanitized AI: your account is ready when it’s done. This checklist takes you from there to a protected team.

  1. Sign in to the dashboard.

    Go to app.sanitized.ai, enter your work email, and click Continue. Then click Email me a sign-in link and open the link from your inbox, or sign in with your company account. For details, see Signing in.

  2. Run the guided setup.

    After you sign in, a panel titled Finish protecting your organization appears in the corner of the dashboard. When it’s collapsed, it shows as a Finish setup button. Click Resume to open the guided setup. It asks one question per screen. Click Save and continue to move on, or Back to change an answer. Your progress is saved in your browser, so you can leave and come back.

    First, choose your protection setup:

    • Optimal setup (recommended): answer a few questions about the data you handle and the AI tools you have an agreement with, and get a suggested policy.
    • Maximum protection: turn on every guardrail on every supported AI site.
    • Basic protection: protect common personal and financial information, such as emails, phone numbers, and credit cards.
    • Custom setup: pick guardrails and AI sites yourself from the full lists.

    If you choose Optimal setup or Basic protection, a later question asks which AI services to guard. The default, All sites except those with an agreement, leaves the AI tools you have an agreement with unprotected. To protect those tools too, choose All sites.

    Then choose how Sanitized AI responds when it catches something. Pick System Default unless you have a reason not to: it blocks the most sensitive data and warns on the rest. For what each option means, see Enforcement actions.

    The last question asks which language you want to use. It only sets your own dashboard, extension and email language; everyone else keeps English unless you change their language on the Users page.

    Click Complete setup. Your guardrails and AI sites are now on. You can change any of them later, or run the guided setup again from My Organization > General > Restart onboarding.

  3. Optional: point the warning to your AI policy.

    If your organization has an AI usage policy, go to My Organization > Protection, and enter its name and link in AI usage policy. The warning people see links to it. See Organization settings.

  4. Optional: choose how people sign in.

    By default, people sign in with a link sent to their work email. To add Google, Microsoft, or Okta sign-in, go to My Organization > Sign-in & provisioning. See Choose how people sign in.

  5. Add your people.

    On the Users page, click Add User for one person or Import Users for a spreadsheet. Give each person an access level and, if they have a manager in Sanitized AI, a Reports to setting. To prepare everything before anyone gets an email, clear Activate now (or Activate everyone now when you import). See Add and manage people.

  6. Test it yourself.

    Install the browser extension, sign in to it, and type a prompt with a made-up name and phone number, such as Call Jane Doe at (514) 555-0134, into an AI site that the guided setup turned on. If you’re not sure which sites are on, check the Guarded Sites page. A Security Check should appear before the prompt is sent. If you chose Monitor All, no warning appears; on the Events page, turn on Monitor only to see the event. See Install and sign in.

  7. Roll out the extension and activate your people.

    If your IT team manages your browsers, ask them to deploy the extension first. Then activate your people. Each person gets a welcome email that tells them what to do for their access level. See Roll out the extension.

  8. Watch the first week.

    Check the Overview and Events pages every day or two for guardrails people often send anyway and for false-positive reports. On the Users page, the Last Login column shows who hasn’t signed in yet. See Events and Review false-positive reports.

After the rollout