Target Appliers

A target applier is the runtime bridge that lets an approved AI suggestion call an owning module command. It is not a policy by itself. It is the concrete, command-safe path that applies a suggestion after the AI action policy, use authorization, permissions, review state, and resource access checks have allowed it.

The target-applier inventory is visible under AI Governance -> Operations.

Why this matters

An action policy may allow AI to suggest or apply an action, but Novantra still needs a module-approved application path. Without a registered target applier, AI cannot write beyond the built-in metadata application path.

This prevents AI from creating a parallel write path into document, risk, control, evidence, lifecycle, incident, or other governed records.

What the table shows

The Operations tab shows:

the registered applier key;
the target module and resource type, when the applier is scoped;
the target action key, when the applier is action-specific;
whether the applier itself blocks review-bypass application;
the boundary used to apply the suggestion.

For V1, the boundary is owning module command. That means the applier must call the target module’s application command or actioner. It must not write directly to the database or create final owning-module truth inside AI.

Built-in appliers

The AI feature currently registers command-safe creation appliers for the governed modules and change-management workflows most commonly produced by reviewed Document Intelligence and AI watcher suggestions.

Applier	Target	Review posture
`dms.create-draft-version`	Creates a DMS successor draft version through the DMS command seam from a reviewed proposed-version suggestion. The draft starts as an editable DMS version; AI does not write the document body or approve it.	Human review required; review-bypass application is blocked by the applier.
`governed-findings.create-finding`	Creates a governed finding through the Findings module command seam from a reviewed AI suggestion.	Human review required; review-bypass application is blocked by the applier.
`governed-risks.create-risk`	Creates a governed risk through the Risks module command seam from a reviewed AI suggestion.	Human review required; review-bypass application is blocked by the applier.
`governed-controls.create-control`	Creates a governed control through the Controls module command seam from a reviewed AI suggestion.	Human review required; review-bypass application is blocked by the applier.
`governed-obligations.create-obligation`	Creates a governed obligation through the Obligations module command seam from a reviewed AI suggestion.	Human review required; review-bypass application is blocked by the applier.
`governed-evidence.create-requirement`	Creates an evidence requirement through the Evidence module command seam from a reviewed AI suggestion.	Human review required; review-bypass application is blocked by the applier.
`change-management.create-change-request`	Creates a draft change request through the Change Management module command seam from a reviewed proposed-edit or redline suggestion.	Human review required; review-bypass application is blocked by the applier.
`privacy-processing.create-dpia`	Creates a privacy DPIA candidate through the Privacy Processing command seam from a reviewed privacy-review suggestion. It does not complete the assessment or clear the document for use.	Human review required; review-bypass application is blocked by the applier.
`privacy-processing.create-transfer-record`	Creates a draft privacy transfer or export-control review record through the Privacy Processing command seam from a reviewed export-control suggestion. It does not authorize external release.	Human review required; review-bypass application is blocked by the applier.
`governed-retention.create-retention-rule`	Creates a draft governed retention rule through the Governed Retention command seam from a reviewed retention-policy or retention-rule suggestion.	Human review required; review-bypass application is blocked by the applier.
`document-governance.create-requirement`	Creates a draft document-governance requirement through the Document Governance command seam from a reviewed document retention or governance-review suggestion.	Human review required; review-bypass application is blocked by the applier.

These appliers record sanitized AI suggestion, source, command, and resulting owning-module references. They do not copy raw prompts, provider responses, source document text, or unreviewed AI narrative into the owning record.

Applying suggestions

When a suggestion is applied through a target applier:

Novantra loads the reviewed suggestion and action policy.
Use authorization, action policy, permission, resource-access, and review state are evaluated.
The registered target applier is selected by key and target scope.
The applier calls the owning module command.
Novantra records the AI action application with sanitized command and result snapshots.

If any required gate fails, the application is blocked.

Cloud, self-managed, and no-AI deployments

Target appliers live inside the optional AI feature bundle. AI-enabled cloud and self-managed deployments can expose the inventory route and UI. No-AI deployments do not include the route, UI, or package import.