IntelliThreat — Site Header (Fable 1.0)
30-Day Trial
Permissions & Data Handling

Exactly what IntelliThreat AI can and cannot see in your Workspace.

Four OAuth scopes. Two read your audit logs. Two act on a compromised account. None of them grant access to the contents of an email or a file. Every permission is granted in your own Admin console, and you can revoke all 4 there at any time.

30 days free · No credit card required · Plans for teams of 10 to 10,000

4 scopesNo mailbox accessNo file contentsRevocable in your Admin console
What It Reads

Two Google APIs. Nothing Else.

IntelliThreat AI collects Workspace activity through Google's own documented APIs. There is no agent to install, no mail routing to change, and no traffic to redirect.

Admin SDK Reports API

The audit and activity record Google already keeps for your domain — sign-ins, administrative changes, Drive activity, group changes, account security events. This is the detection substrate.

Alert Center API

The alerts Google generates itself: phishing, malware, suspicious sign-in, government-backed attack warnings. IntelliThreat AI enriches each one with the surrounding activity so an alert becomes an incident with a story attached.

A Service Account You Authorize

Access runs through a dedicated service account with domain-wide delegation, authorized by your administrator in the Google Admin console and limited to the 4 scopes below. Setup takes ~10 minutes.

The Four Scopes

The Only 4 Permissions It Asks For

ScopeWhat it doesWhat it does not grant
admin.reports.audit.readonlyRead-only access to audit and activity logs. Powers detection.No write access of any kind. No message or document contents.
apps.alertsRetrieves alerts from Google's alert center. Powers detection.No ability to create, suppress, or alter Google's own alerts.
admin.directory.userResponse actions on user accounts, such as suspending a compromised account.No access to mailbox contents. No access to Drive file contents.
admin.directory.user.securityResponse actions such as signing a user out of every session and revoking third-party app access.No password reading. No session content. No file contents.

The first 2 scopes detect. The last 2 respond. None of the 4 grant access to the contents of an email or a document.

Read-Only Until You Say Otherwise

The 2 response scopes exist so IntelliThreat AI can contain a compromised account in seconds instead of waiting for someone to wake up. Whether it acts on its own, or asks first, is your configuration — not a default you inherit.

What It Never Collects

It Does Not Read Your Mail. It Does Not Read Your Files.

IntelliThreat AI collects activity and event logs: who signed in, from where, what changed in the Admin console, what moved in Drive, which groups were altered, which account security settings were touched. It does not access mailboxes, and it does not store the contents of your documents.

What Log Records Do Contain

Investigation requires some descriptive detail: file names, sharing settings, and user email addresses appear in log records, because an exfiltration incident cannot be explained without them.

The One Exception, Stated Plainly

When Google's alert center flags a specific message — as phishing or malware — Google includes that message's subject line and a short excerpt in the alert it generates. IntelliThreat AI stores those details with the alert so an analyst can see what Google flagged. That excerpt comes from Google's alert, not from reading the mailbox, and it is protected by the same controls as every other record.

No mailbox access. No document contents. The only message text ever stored is the excerpt Google itself puts inside an alert.

Where the Data Lives

Two Kinds of Data. Both Isolated to You.

Activity Logs and Alerts

Stored in indexes dedicated to your organization, inside storage infrastructure dedicated to Google Workspace data. Every request passes through the platform's authorization layer, so a user can only ever retrieve data for an organization they belong to. Data is encrypted in transit and at rest.

AI Behavior Profiles

To recognize what normal looks like in your environment, IntelliThreat AI maintains behavior profiles — typical sign-in locations, working hours, usual sharing patterns. These are held in a separate AI knowledge store with database-enforced tenant isolation: row-level security policies, per-tenant encryption keys, and physical shard isolation per tenant.

Your organization's behavior data is never mixed with another organization's, and never used to inform analysis of another organization's environment.

The AI

Your Logs Never Reach a Third-Party Model.

IntelliThreat AI runs on its own in-house large language model. Workspace logs, alert data, and any contextual information used in analysis are never sent to OpenAI, Google, Microsoft, or any other third-party AI platform. Analysis happens inside IntelliThreat's controlled environment, from collection through to the alert that lands in front of your team.

Platform Access Controls

Controls Inside the Platform Itself

Role-Based Access Control

Access is governed by role. Team members see what their role requires and nothing beyond it.

Multi-Factor Authentication, Mandatory

Required for every user. Not an option, not a recommendation. Sensitive administrative actions require re-verification on top of it.

Tamper-Evident Audit Trail

Every administrative action taken in the platform is recorded in an audit trail built to show evidence of alteration.

Scoped Access and Retention

Only the permissions the platform needs, and none that reach mailbox or file contents. Log data is retained only as long as delivering the service requires. The full retention policy is available on request.

No Resale, No Sharing

Your data is never sold, never shared with third parties, and never used for any purpose other than delivering the service to you.

Standards

Built Around NIST 800-171 and the OWASP Top 10

The platform was designed around the security requirements of NIST SP 800-171 and the OWASP Top 10 — mandatory multi-factor authentication, role-based access control, re-verification for sensitive actions, tamper-evident audit logging, and strict separation between client organizations.

IntelliThreat does not currently hold a third-party compliance certification. Security questionnaires are completed on request, and the team will walk through these controls with yours directly.

FAQ

Questions your team will ask.

What permissions does IntelliThreat AI need in our Google Workspace?
A service account with domain-wide delegation, authorized for 4 scopes: admin.reports.audit.readonly and apps.alerts for detection, admin.directory.user and admin.directory.user.security for response. None of the 4 grant access to email contents or file contents. You authorize them in your Admin console and can revoke them there at any time.
Does the platform store the actual content of emails or files?
No. It collects activity and event logs — sign-ins, administrative changes, Drive activity, group changes, account security events. It does not access mailboxes and does not store document contents. Log records do include descriptive details needed for investigation, such as file names, sharing settings, and user email addresses.
Does our data get sent to a third-party AI provider like OpenAI or Google?
No. IntelliThreat AI runs on its own in-house large language model. Workspace log data, alert information, and client-specific context are never sent to an external AI service. Data stays inside IntelliThreat's controlled infrastructure from collection through to alert delivery.
Is our data isolated from other organizations?
Yes, at two layers. Activity logs and alerts sit in indexes dedicated to your organization behind an authorization layer that scopes every request to the organizations a user belongs to. AI behavior profiles sit in a separate store with row-level security policies, per-tenant encryption keys, and physical shard isolation per tenant.
What security standards is the platform built around?
NIST SP 800-171 and the OWASP Top 10 — mandatory multi-factor authentication, role-based access control, re-verification for sensitive actions, tamper-evident audit logging, and strict separation between client organizations. IntelliThreat does not currently hold a third-party compliance certification, and completes security questionnaires on request.
Can we revoke access after deployment?
Yes, entirely, from your Google Admin console, without involving IntelliThreat. Revoking the 2 detection scopes stops collection. Revoking the 2 response scopes leaves detection running and removes the platform's ability to contain an account.
You Hold the Keys

Granted in Your Admin Console. Revoked There Too.

Every one of the 4 scopes is authorized by your administrator, in your Google Admin console. You can review them, narrow them, or revoke them entirely, at any time, without involving IntelliThreat. Revoking the detection scopes stops collection. Revoking the response scopes leaves detection running and takes containment off the table.

PDF Download

Your download is ready

 

You’ll be redirected straight to the PDF. We never share your info.

Product