Agentic Automation

Agentic automation uses AI agents that plan and carry out multi-step work, calling tools, applications and other agents, inside governed automated processes alongside robots, workflows and people.

RPA follows fixed scripts that imitate user actions. Agentic automation lets an AI agent decide steps at run time, so it can handle variable inputs and goals, usually with orchestration and approval controls around it. See the RPA Software and Intelligent Automation pages.

Agentic AI is the broader field of autonomous, goal-directed AI systems. Agentic automation applies agents to business processes and typically adds orchestration, integration with enterprise systems, auditability and human oversight. See the Agentic AI page.

The Model Context Protocol (MCP) connects AI applications to tools and data, and the Agent2Agent (A2A) protocol lets agents communicate with each other. Both are open specifications. See the Agentic AI Communication Protocols page.

Vendors describe governance features such as access control, audit logs, guardrails and human approval steps; check each product's documentation. See the Guardrails for Agentic AI and AI Agent Governance Solutions pages.

Usually with a well-understood process, discovered or modeled first, then adding agents to steps that need judgment while keeping deterministic steps in workflows or RPA. See the Process Discovery and Workflow Engines pages.

Links were checked on 31 October 2026. Release facts come from GitHub, and descriptions follow each source's own pages; they are not rankings. A few vendor product pages that returned 404 were not used.