Integrations
Bring supported documents and AI-use signals into a review, within the permissions you grant. Each connection contributes something specific; a signal is a starting point, not a verdict.
Core connections to discover AI, review changes and turn findings into team work.
Finds AI SDKs, coding agents and MCP servers in your repositories and watches every change.
Read-only metadata and contents of the repositories you choose. Fencia reads known paths (manifests, lockfiles, agent configuration, workflows), never secrets or .env files.
Analyses the default branch. A dependency proves code exists, not that it is deployed.
Shows which AI apps your people authorised with their work account.
Directory read with an administrator's consent. Stores counts, never who.
Connection and directory reads verified in our QA tenant; acceptance in your tenant is still required. Authorising an app does not mean its use is approved.
Shows which AI apps your people authorised with their Google Workspace account.
Administrator audit reports (token activity), read-only. Stores counts, never who.
Not yet verified with a live tenant. Only sees apps connected through OAuth.
Inventories AI resources deployed in Azure and their configuration.
Entra administrator approval and a Reader role assigned by the customer; no access to keys or data.
Beta: built to Azure's documentation and tested with simulated responses; not yet verified with a real tenant.
Inventories AI resources in AWS and their logging configuration.
Read-only cross-account role with an ExternalId; no static keys.
Beta: built to AWS's documentation and tested with simulated responses; not yet verified with a real account.
Detects where Claude Code is used, what its tools and MCP servers may do and whether its configuration changed; and, with hooks, records and can enforce your policy while it works.
Reads configuration committed to connected repositories. Hooks send the full event over HTTPS; Fencia reduces it in memory and retains only metadata, never prompts, contents or arguments.
Beta: hooks were tested against Claude Code's documentation, not a live session. If Fencia is unreachable Claude Code carries on (hard rules belong in its permission lists).
Turns an approved remediation into an issue with evidence, owner and plan.
Write access limited to the chosen project; every issue is created by a person after reviewing it.
Beta: tested with simulated responses; not yet verified with a real Jira site.
Live product
The minimum plan comes from the current product configuration. Beta sources (GitHub, Google Workspace, Microsoft 365) have not yet been verified against a live customer tenant.
Swipe the table to compare every capability
| Source and scope | Minimum plan | Document import | AI-use signals | Scheduled document checks |
|---|---|---|---|---|
| Google DriveDocuments available through the connection. File names and supported document text can also contribute AI-use signals. | Starter | Included | Included | Included |
| OneDriveDocuments available through the Microsoft connection. This source is not part of the AI-use signal sweep. | Team | Included | Not included | Included |
| SharePointDocuments available through the Microsoft connection. This source is not part of the AI-use signal sweep. | Team | Included | Not included | Included |
| DropboxDocuments available through the connection. This source is not part of the AI-use signal sweep. | Team | Included | Not included | Included |
| SlackMessages from public channels the connected bot can access, within the previous seven days. Slack can also receive notifications. | Team | Not included | Included | Not included |
| Google WorkspaceBeta. The AI apps people authorised with their Google account, from the administrator's token activity: how many people, and what each app can read. Never mail or files. Needs a Workspace administrator. | Team | Not included | Included | Not included |
| Microsoft 365Beta. The AI apps with delegated permissions in Microsoft 365 (Entra ID): how many people, and what each app can read. Never mail or files. Needs a Microsoft 365 administrator. | Team | Not included | Included | Not included |
| GitHubAI-use signals from dependency manifests in recently updated repositories. GitHub does not import documents or run scheduled patrols; provider validation is still pending. | Team | Not included | Included | Not included |
| Jira | Team | Not included | Not included | Not included |
| Azure | Team | Not included | Not included | Not included |
| AWS | Team | Not included | Not included | Not included |
Connection access depends on your plan and granted permissions. A signal starts a review; your team decides what it means for a policy or legal obligation.
Get started