Skip to main content

Agentic Automation on Automation Anywhere

Automation Anywhere — long known for its RPA platform — now positions itself as a provider of agentic automation, built around a product called AI Agent Studio.

What AI Agent Studio is​

AI Agent Studio lets you build goal-driven AI agents that automate work a fixed RPA bot can't handle on its own: reading a document and deciding what it means, triaging a support ticket, checking whether a case follows a compliance policy, or recommending a next action. Agents built here are powered by the Process Reasoning Engine (PRE), which handles the reasoning, decides when a human needs to be looped in, and orchestrates across bots, APIs, documents, and even other agents in real time.

Rather than replacing Automation Anywhere's existing RPA bots, agents built in AI Agent Studio typically sit on top of them — an agent can decide what needs to happen next and then call an existing bot as one of its tools to actually execute the steps in a legacy system.

Who builds with it​

AI Agent Studio is designed to be usable on day one without a long IT implementation. It offers low-code tools so both professional developers and citizen (business) users can select a foundation model, connect enterprise data sources, write and optimize prompts, and link them to specific actions — typically producing a first working version of an agent in a matter of days.

Where it fits in the platform​

  • AI Agent Studio — where you design and build agents (this is where the tutorials in this section focus).
  • Control Room / Automation Co-Pilot — where agents and bots are deployed, monitored, and governed.
  • AI Guardrails — Automation Anywhere's built-in safety and governance layer, which blocks unsafe agent actions before they happen and keeps behavior aligned with your policies.

Getting access​

Automation Anywhere offers a free Cloud Community Edition you can use to start experimenting with AI Agent Studio without a paid license — this is the fastest way to follow along with the walkthrough in this section.

Next: Core concepts covers the vocabulary you'll need — agents, skills, and the Process Reasoning Engine — before you build anything.