Help Center
DocsAI Agents Overview
Autonomous AI agents that proactively manage your CRM, producing action plans for your review
What Are AI Agents?
AI Agents are autonomous workers that analyze your CRM data and propose actions on a schedule.
AI Agents work in the background to keep your CRM healthy and your pipeline moving. Unlike the AI Assistant (which responds to your questions), agents proactively:
- Analyze your CRM data on a schedule or when triggered
- Identify items needing attention (stale deals, missing data, overdue tasks)
- Propose specific actions with confidence scores and reasoning
- Execute approved actions or auto-execute low-risk ones
By default, all agent actions go through a review queue ("Ask me first" mode). You can switch to "Go ahead" mode to auto-execute actions. Per-tool permission overrides give you fine-grained control over which tools can act autonomously.
AI Agents vs. AI Assistant
Two complementary tools for different use cases.
| Feature | AI Assistant | AI Agents |
|---|---|---|
| Trigger | You ask a question | Scheduled, manual, or event-based |
| Behavior | Reactive, conversational | Proactive, autonomous |
| Output | Chat responses | Action plans for review |
| Interface | Chat panel | Agent pages + review queue |
Data Access & Security
Agents respect the same access controls as your CRM account.
Each agent runs with the permissions of its creator. A sales user's agent only sees their assigned records, while a manager's agent can see all organization data.
Agents cannot access other organizations' data, system internals, or any information beyond what's available in the CRM interface.
Agent runs are subject to your organization's billing limits. If your trial has expired, your AI token limit is reached, or your personal allocation is exhausted, agent runs will fail with a descriptive message. Check Settings > Usage to monitor your limits.