OpenAI Agents on UiPath: Building Multi-Agent Systems

[Virtual Hands on workshop]
Modern AI can generate text, answer questions, and assist with tasks but real-world enterprise problems are rarely solved with a single prompt. In this session, we’ll explore how agentic systems take AI beyond simple interactions by enabling multiple specialized agents to collaborate, reason, and execute complex workflows.

📅What the day looks like?

• The Limits of Single-Prompt AI
  We begin by exploring why single prompts and traditional workflows struggle with complex enterprise tasks.
  Understand the challenges of multi-step reasoning, decision-making, and automation when relying only on prompt-based interactions.
• When Agent Logic Needs to Be Coded
  As systems grow more complex, prompts alone are not enough.
  This section discusses when and why developers need structured agent logic, and the limitations of purely prompt-driven or low-code approaches for advanced agentic workflows.
• Understanding the OpenAI Agents Architecture
  Dive into the core building blocks of agent systems, including:
  Agents, Tools, Guardrails, Sessions, The Agent Execution Loop
• Designing Multi-Agent Systems
  Discover how specialized agents collaborate as a team to solve complex problems.
  We’ll explore routing, delegation, and handoff patterns that allow agents to coordinate and complete multi-step tasks efficiently.
• Extending Agents with the UiPath SDK
  See how the UiPath SDK for OpenAI Agents enhances agent capabilities with enterprise-ready features and integrations across the UiPath automation ecosystem.
• Live Demo & Hands-On Walkthrough
  Watch and build a multi-agent system using the UiPath OpenAI Agents SDK.

❗Prerequisites

• Join from your laptop
• Python 3.10+
• Code editor (VS Code or equivalent)
• UiPath Automation Cloud account

🎯Good to Have

• Basic Python scripting knowledge
• Understanding of Agent fundamentals

Speaker: Rajneesh Khare [Automation Developer, WonderBotz]