---
title: Control enforcement points
description: The systems an AI Control Tower policy can act on.
locale: en-US
canonical_url: https://www.servicenow.com/docs/r/australia/intelligent-experiences/gov-pol-enforcement-points.html
release: australia
topic_type: reference
last_updated: "2026-08-27"
reading_time_minutes: 1
keywords: [ServiceNow Otto, AI Agents, generative AI, agentic AI, Policies, enforcement points, connectors]
breadcrumb: [Reference, Controlling AI asset usage, Govern AI assets, AI Control Tower, Enable AI experiences]
---

# Control enforcement points

The systems an AI Control Tower policy can act on.

<table id="table_pol_enforcement_points"><thead><tr><th>

Enforcement point

</th><th>

Description

</th></tr></thead><tbody><tr><td>

**Agent Client Collector \(ACC\)**

</td><td>

A lightweight agent installed on employee devices. Blocks and unblocks AI agents by model or domain, enforced at the device regardless of which application or browser someone uses to reach it.

</td></tr><tr><td>

**AWS Bedrock**

</td><td>

Blocks and unblocks Amazon Bedrock agents.

</td></tr><tr><td>

**AWS Bedrock AgentCore**

</td><td>

Blocks and unblocks Amazon Bedrock AgentCore agents.

</td></tr><tr><td>

**Azure AI Foundry**

</td><td>

Blocks and unblocks Azure AI Foundry agents.

</td></tr><tr><td>

**Gemini Enterprise Agent Platform agents**

</td><td>

Blocks and unblocks Gemini Enterprise agents.**Note:** Applies to agents with unique identities only.

</td></tr><tr><td>

**Okta**

</td><td>

Blocks and unblocks a person's access through their Okta-managed identity.

</td></tr><tr><td>

**ServiceNow agents**

</td><td>

Blocks and unblocks AI agents on the ServiceNow AI Platform®.

</td></tr></tbody>
</table>**Parent Topic:**[AI Control Tower policies reference](https://raw.githubusercontent.com/ServiceNow/ServiceNowDocs/australia/markdown/australia/intelligent-experiences/gov-pol-reference.md)