Agentic Automation on UiPath
UiPath — one of the largest RPA vendors — describes itself as offering the first enterprise-grade platform for agentic automation, built around two complementary products: Agent Builder and Maestro.
What Agent Builder and Maestro are
Agent Builder is where you create individual AI agents: you define a system prompt, give the agent tools it's allowed to call, ground it in business context, and configure escalations for when it needs a human. Agent Builder supports two paths — a low-code visual canvas for prototyping quickly, or a code-first Python SDK for teams that want deeper customization — and both paths lead to production-ready agents.
Maestro is the orchestration layer above individual agents. It's a cloud-native, low-code BPMN (Business Process Model and Notation) builder that lets teams design end-to-end business processes visually, coordinating multiple agents, existing RPA robots, and human steps within a single, unambiguous process definition. Maestro is what gives an enterprise centralized oversight as it scales from one agent to many.
How agents relate to UiPath's existing RPA robots
Agents built with Agent Builder aren't a replacement for UiPath's existing Robots (the traditional RPA building block) — they're designed to call them. Maestro can integrate agents built via Agent Builder (or even external LLM-based agents) into the same orchestrated workflow as existing robots, managing fallback logic and human escalation across the whole process.
Where it fits in the platform
- Agent Builder (Studio Web) — where individual agents are designed, tested, and published (this section's tutorials focus here).
- Maestro — where multiple agents, robots, and human steps are orchestrated into a full business process.
- Orchestrator — where published agents and robots are deployed, managed, and monitored.
Getting access
UiPath offers a free tier with no credit card required, which is enough to follow the walkthrough in this section.
Next: Core concepts introduces the vocabulary — tools, context grounding, and escalations — that Agent Builder is built around.