Agentic Tasks bring AI-powered collaboration directly into your Cutover runbooks, enabling users to interact with an AI assistant from within a task. Whether you're investigating incidents, gathering information, executing approved actions, or automating operational workflows, Agentic Tasks provide a seamless way to combine human decision-making with AI assistance.

Traditionally, Cutover integrations are designed to perform a specific action or exchange data with an external system. An integration executes a predefined workflow and returns a result.
Agents introduce a different approach.
Rather than executing a single predefined action, an agent can interpret requests, maintain context, make decisions, and interact with users throughout a task. Agents operate as active participants within a runbook, enabling more dynamic and conversational workflows.
In Cutover, agents are surfaced through dedicated task types, allowing users to engage with AI-powered capabilities directly within the runbook experience.
Agentic Tasks extend Cutover workflows with AI-powered automation while keeping the runbook as the system of record. Conversations, approvals, actions, and outcomes are captured within Cutover, ensuring visibility, governance, and a complete audit trail. While agents can interact with external systems, all activity is surfaced back into the runbook so operators maintain oversight and control throughout execution.
Agentic Tasks are the primary way users interact with agents in Cutover.
When an Agentic Task starts, users can communicate with the agent directly through a dedicated chat panel interface within the task. Conversations are persistent, shared with other users who have access to the task, and remain available throughout the task lifecycle.
Depending on the configured AI capability, the agent can:
For actions that require human oversight, Agentic Tasks support approval workflows, ensuring users remain in control of important decisions. Approval decisions and resulting actions are recorded within the runbook to provide a clear audit trail.
Agentic Tasks can be used for a variety of operational activities, including:
- Incident investigation and triage
- Log analysis and troubleshooting
- Change validation
- Automated operational checks
- Generating summaries and reports
Cutover AI Assistant Chat is a built-in AI assistant that requires no additional configuration. The assistant automatically understands the task and runbook context, allowing users to ask questions and perform supported actions directly within a runbook.
Note: The Cutover AI Assistant is enabled on request. Please contact your Customer Success Manager (CSM) to discuss access.
Organisations can also connect their own AI agents using Cutover integrations. Custom agents communicate with Cutover and can be tailored to specific operational, business, or technical use cases.
Agentic Tasks provide:
Agentic Tasks operate within Cutover's existing permission model. Users can only perform actions that they are authorised to perform, and all approval decisions and AI-assisted actions can be audited through standard platform governance controls.
When integrating custom AI agents, we recommend following least-privilege principles and implementing appropriate authentication, approval, and validation mechanisms. Custom agents cannot perform actions within Cutover unless they are instrumented with Cutover MPC.
Choose the guide that best matches your role:
Learn how to use Agentic Tasks within a runbook, interact with the AI assistant, manage conversations, and approve actions.
See: Agentic Task Type User Guide
Learn how to build and integrate custom AI agents with Cutover, including AG-UI requests, Server-Sent Events (SSE), authentication, context passing, and human-in-the-loop workflows.
See: Develop AI Agents for Agentic Tasks
Watch our video to learn more about Agentic Tasks.
