Agentic Intro and Hands-on: Build your first Coded Agent

2 Statements we make in this presentation may include statements which are not historical facts and are considered forward-looking within the meaning of the Private Securities Litigation Reform Act of 1995, which are usually identified by the use of words such as “anticipates,” “believes,” “estimates,” “expects,” “intends,” “may,” “plans,” “possible,” “projects,” “outlook,” “seeks,” “should,” “will,” and variations of such words or similar expressions, including the negatives of these words or similar expressions. We intend these forward-looking statements to be covered by the safe harbor provisions for forward-looking statements contained in Section 27A of the Securities Act of 1933, as amended, and Section 21E of the Securities Exchange Act of 1934, as amended, and are making this statement for purposes of complying with those safe harbor provisions. These forward-looking statements include, but are not limited to, statements regarding: our ability to drive and accelerate future growth and operational efficiency and grow our platform, product offerings, and market opportunity; our business strategy; plans and objectives of management for future operations; the estimated addressable market opportunity for our platform and the growth of the enterprise automation market; the success of our platform and new releases including the incorporation of AI; the success of our collaborations with third parties; our customers’ behaviors and potential automation spend; and details of UiPath’s stock repurchase program. Forward-looking statements involve known and unknown risks, uncertainties, and other factors that may cause our actual results, performance, or achievements to be materially different from any future results, performance, or achievements expressed or implied by the forward-looking statements. These risks include, but are not limited to, risks and uncertainties related to: our expectations regarding our revenue, annualized renewal run-rate (ARR), expenses, and other operating results; our ability to effectively manage our growth and achieve or sustain profitability; our ability to acquire new customers and successfully retain existing customers; the ability of the UiPath Platform to satisfy and adapt to customer demands and our ability to increase its adoption; our ability to grow our platform and release new functionality in a timely manner; future investments in our business, our anticipated capital expenditures, and our estimates regarding our capital requirements; the costs and success of our marketing efforts and our ability to evolve and enhance our brand; our growth strategies; the estimated addressable market opportunity for our platform and for automation in general; our reliance on key personnel and our ability to attract, integrate, and retain highly-qualified personnel and execute management transitions; our ability to obtain, maintain, and enforce our intellectual property rights and any costs associated therewith; the effect of significant events with macroeconomic impacts, including but not limited to military conflicts and other changes in geopolitical relationships and inflationary cost trends, on our business, industry, and the global economy; our reliance on third-party providers of cloud-based infrastructure; our ability to compete effectively with existing competitors and new market entrants, including new, potentially disruptive technologies; the size and growth rates of the markets in which we compete; and the price volatility of our Class A common stock. Further information on risks that could cause actual results to differ materially from our guidance and other forward-looking statements can be found in our Annual Report on Form 10-K for the fiscal year ended January 31, 2025 filed with the United States Securities and Exchange Commission (SEC), in our Quarterly Reports on Form 10-Q filed with the SEC, and in other filings and reports that we may file from time to time with the SEC. Any forward-looking statements contained in this presentation are based on assumptions that we believe to be reasonable as of this date. Except as required by law, we assume no obligation to update these forward-looking statements. Our fiscal year end is January 31, and our fiscal quarters end on April 30, July 31, and October 31. All third-party trademarks, including names, logos and brands, referenced by us in this presentation are property of their respective owners. All references to third-party trademarks are for identification purposes only. Such use should not be construed as an endorsement of the products or services of us. Safe Harbor

WhereUiPathCommunitycomestogether

Community in a year 1M+ learning completions on Academy 5M Forum pageviews every month; 400+ use cases, use cases and tutorials created by community; 360 MVPs, champions and experts ready to help; 500+ developer meetups across 86 countries The cohesive and friendly-environment of the community allow all to share knowledge and support each other to learn from one another. This diverse and open connection within the community is unbeatable. Alp Uguray – Solutions at Ashling Partners and UiPath MVP for 5 years in a row

JoinNOW: UiPathCommunity Cluj-Napoca Chapter

UiPathForum:AplaceforUiPathdevelopers togethelpandstayupdated!

UiPath Insider Program As an Insider, you get access to early builds of new features and updates—an exclusive opportunity to give feedback on our latest innovations. UiPath Community Blog

The UiPath Most Valuable Professional (MVP) Award is our way of recognizing passionate automation experts who go above and beyond, not just with their UiPath skills, but also in how they support others. More than tech experts, they’re mentors, leaders, and changemakers in the UiPath Community. UiPath Most Valuable Professionals Every year, we welcome a new group of MVPs from around the world, selected for their meaningful contributions and impact. 15 Oct – 20 Nov 2025 Applications open

Community Certification Framework2025 22K registrants sofar 11certification tracksopen!

Build your first Coded Agent The UiPath word mark, logos, and robots are registered trademarks owned by UiPath, Inc. and its affiliates. UiPath (R) is a registered trademark in the United States and several countries across the globe. See TMEP 906. ©2025 UiPath. All rights reserved. Enterprise grade Agents with UiPath

11 Change has never been this fast . And yet it will never be this slow again. of CxOs believe GenAI will be transformative for their company ACCENTURE 97% of execs say fewer than 30% of initiatives will scale in the next 6 months 67% DELOITTE of execs say that AI/ML is driving revenue growth and gross margins now 25% BCG

12 12 12 Average large organization: 175+ ENTERPRISE APPLICATIONS

13 Process is the lifeblood of any organization Yet every process has its challenges Poor experiences Insufficient throughput Process delays Error-prone Siloed data & lack of visibility Resolution & close Cross-team collaboration Ingestion & triage Complex data inputs Research & analysis Cognitive work Data entry Rote, manual steps Approved? Rework loop Index and review documents Data lookup & validation Research similar cases Submit approval Update system Analyze & review Email

You have options… But the power is in the whole, not the parts. OPTION 1: POINT SOLUTIONS OPTION 2: ENTERPRISE-GRADE AGENTIC AUTOMATION ▪ Longer time-to-market ▪ Increased TCO ▪ Fragmented visibility App-specific automation CRM workflow Custom Code Python Agent In-app configuration Home-grown App ▪ Unified platform ▪ Enterprise governance & security ▪ End-to-end agentic automation Observability & Governance Code-tracing, Centralized Logging, Dashboarding Triage email Index & review documents Research similar cases Review Update system UiPath end-to-end orchestrated process “It’s about how the pieces fit together, not the pieces themselves. The challenge lies in orchestrating the range of interactions and integrations at scale. Each use case often needs to access multiple models, vector databases, prompt libraries, and applications.”

15 There’s a solution to every challenge Agentic AI Human-in-the-loop Robotic Process Automation (RPA) Intelligent Document Processing (IDP) Resolution & close Cross-team collaboration Ingestion & triage Complex data inputs Research & analysis Cognitive work Data entry Rote, manual steps Approved? Rework loop Index and review documents Data lookup & validation Research similar cases Submit approval Update system Analyze & review Email AGENTICORCHESTRATION

UiPath journey to agentic automation 2018 RPA Computer Vision API Integration Low- code Apps Orchestration Governance Cloud Process Discovery Analytics Continuous Testing Intelligent Document & Communication Processing GenAI Automation UiPath LLMs UiPath GenAI Activities 2020 2022 2023 2024 2025 Agents Robots People HUMAN-IN-THE-LOOP (HITL) INTELLIGENT DOCUMENT PROCESSING (IDP) AGENTIC AI Email Index and review documents Data lookup & validation Research similar cases Update system Analyze & review Submit approval Approved? Rework loop Email Index and review documents Research similar cases Update system Analyze & review Submit approval ROBOTIC PROCESS AUTOMATION (RPA) END-TO-END PROCESS ORCHESTRATION AND AUTOMATION Email Index and review documents Research similar cases Update system Analyze & review Submit approval Approved? Rework loop Email Index and review documents Data lookup & validation Research similar cases Update system Analyze & review Submit approval Research similar cases

Until recently, you’ve focused on orchestrating robots and people. ARE YOU READY TO ADD AI AGENTS TO THE MIX?

Enterprise Agents Trustworthy AI agents that act independently and make dynamic decisions Agents think. Robots do. People lead.

Tools Automations APIs Other Agents Escalation HITL Conversational Context Enterprise Knowledge Documentation Refinement Execution Planning Prompts Models End Tools Automations APIs Agents Escalation HITL Conversational Context Enterprise Knowledge Documentation Refinement Execution Planning Prompts Models End Start Agents • Autonomous, Conversational • Tools, including MCP, multiple integrations via Integration Service apps • Contexts (RAG, DeepRAG) • Escalations (via Action Center) Security and Reliability • Guardrails • Evals • Agent score • Autopilot • AI Trust Layer • Maestro – Agentic Orchestration

Need to supercharge the agents with full coding control and flexibility?

Coded agents overview 23 What is it? Code fully integrated as an Agent Process in UiPath How does it work? • Native compatibility with agent framework (LangChain, LlamaIndex) • Integrated via the UiPath Python SDK to access platform services such as assets, buckets, context grounding, and the LLM Gateway • Packaged and published using the CLI as a native UiPath Package, then deployed as a governed, secure Process on the UiPath Platform Pack and publish using UiPath CLI UiPath assets, context grounding, LLM Gateway via UiPath Python SDK

5 Get started Tailor it Gain trust Make it UiPath-native Deploy run, integrate, monitor, troubleshoot Start in IDE • Create project from scratch or from a sample/template • Use frameworks like LlamaIndex or LangGraph 24 Customize agent behavior • Adjust logic and agent behavior to match desired output • Integrate with UiPath assets, buckets, context grounding, abilities. • If you can code it, you can do it. Test and evaluate • Run and refine the agent until it produces the desired output • Define and run evals such to ensure it doe s what it should, and nothing it shouldn’t. Pack and publish to UiPath • Use the UiPath CLI SDK to package the agent • Publish it as a UiPath- native artifact Orchestrate in UiPath • Leverages the established Package -> Process -> Job lifecycle • Real-time logs, traces, drift detection, cost alerts • Leverages AI Trust Layer for LLM Calls • Integrate with larger workflows and other processes. Journey to production 1 2 3 4

CodedAgents in the grand scheme WherecodedagentsfitintotheoverallUiPathvision Deliver on the UiPath mandate • The platform of choice for Agentic Automation —open, flexible, and securely governed. • As first-class artifacts in the UiPath Platform, they inherit all governance, security, and orchestration capabilities, ensuring agentic automations are built and managed with enterprise-grade confidence. Feature parity across approaches • Including Traces, AI Trust Layer, Evals, Scoring, MCP, and beyond. • Operationally, low code and coded agents are treated equally in Orchestrator, Agent Instance Management, Maestro, Solutions, AI Trust layer… Coherent authoring and deploy experience • One can start with low-code, convert to code, and extend the logic as needed. • With LLM-powered coding lowering the barrier to entry, the authoring gap between low-code and coded experiences is steadily diminishing 25

Why CodedAgents? 26 CONTROL • Direct control over LLM calls • Deterministic logic with flexible coding • Blend coded steps with LLM-powered decisions FLEXIBILITY • “If you can code it, you can do it” • Adapt logic to any requirement • Push boundaries of agent capabilities COMPLEXITY • Branching agent logic beyond simple calls • Multiple nodes, steps, and trajectories • Supports nuanced decision-making ProDev LIFECYCLE • Inherit versioning, testing, debugging practices • Leverage open-source SDK integration • Fits into standard professional dev workflows REUSABILITY. EXTENSIBILITY • Reuse agents across multiple scenarios • Iteratively evolve agents for accuracy & new capabilities CUSTOMIZATION • Build niche or specialized scenarios • Integrate with bespoke software solutions • Create & use coded MCP Servers

Open and flexible SDK Open & flexible SDK No lock-in Author with the IDE of choice Open Source SDK Interoperability (MCP/A2A) Native Compatibility with Agent framework LLM Agnostic No lock-in. Framework, LLM Provider and Interoperability agnostic.

Why CodedAgents? 28 CONTROL • Blend coded steps with LLM-powered decisions • Direct control over LLM calls FLEXIBILITY • Push boundaries of agent capabilities • “If you can code it, you can do it” • Adapt logic to any requirement COMPLEXITY • Branching agent logic beyond simple calls • Multiple nodes, steps, and trajectories • Supports nuanced decision-making ProDev LIFECYCLE • Inherit versioning, testing, debugging practices • Leverage open-source SDK integration • Fits into standard professional dev workflows REUSABILITY. EXTENSIBILITY • Reuse agents across multiple scenarios • Iteratively evolve agents for accuracy & new capabilities CUSTOMIZATION • Build niche or specialized scenarios • Integrate with bespoke software solutions • Create & use coded MCP Servers

Decision tree between low-code vs. code Comfortable using code? Start with low-code • You're a citizen developer or business user who wants to move fast • You prefer designing via drag and drop • Agent does not require heavy customization • See business value quickly and confidence to expand further Start with code authoring • You're a professional developer (including RPA devs) • The project requires advanced or custom functionality • You need granular control over the agent's logic Low-code for speed and scoring, code for control • You’re a developer who wants to prototype fast, then refine with code • You need a quick way to measure, score or demo before committing fully • You like to vibe code – experiment visually first then go to code for precision Starting a new agentic automation project No Yes Yes, but…

Use anywhere. Register all types of MCP servers in UiPath and call them from UiPath agents or external LLM agents — same skills, same policies. 30 Expand agent toolsets with MCP Safer skill reuse. Give agents new skills, reuse them across processes, and test them quickly before going live. Standard tool protocol. MCP lets agents/LLMs call services without custom adapters.

UiPath Controlled Agency via layered scope of control How UiPath bends the curve the power of AI Agents operating within clear guardrails leveraging industry-leading UI and API automation to complete your process in the right way for you Appropriate supervision AI Trust Layer governance Ex) Model config rules, PII filters, audit logs All agents Ex) execution thresholds, deployment governance Agent components Ex) OR+ policies, Action Center preferences Individual agents Ex) Tool guardrails, agent memory/traces Model selection policies Dynamic evaluations Agent automation ops policies High-quality context grounding Workflows as tools Prompt auto-tuning Execution audit trails

What’s next? Aglimpseintowhatwe’llbereleasing Build • Studio Web integration (Run - Debug - Publish) • Coding Agents support (.md specs for Codex, Cursor, etc.) Evaluate • Coded Evals • Studio Web Evals Run • Coded Agents in Automation Suite Connect • Integration Service Activities as MCP tools • MCP going GA with full support in Studio Web, Low Code agents and Solutions • A2A support 33 Conversational CodedAgents

Wrap-up – What you can do next UiPath:Enterprise-gradeCodedAgents,end-to-end Build • Start from the sample agent, connect tools, use Orchestrator RAG • See other samples • Start from scratch – you can do it, coding agents can help Operationalize • Package using UiPath CLI • Add evals in SW • Deploy as a process • Monitor in Orchestrator, Agent Instance monitor Scale • Include it + orchestrate it in Maestro • Use escalations where it matters via Action Center • Audit + Policies + PII protection via AI Trust Layer Join our next sessions • November 27th UiPath LangGraph Agent. Build and integrate an enterprise ready Agent with UiPath • December 18th Gain trust in your Coded Agent before moving to production with UiPath 36

Thank You. …and see you in the next UiPath Community sessions!!!

Agentic Intro and Hands-on: Build your first Coded Agent

More Related Content

Similar to Agentic Intro and Hands-on: Build your first Coded Agent

More from UiPathCommunity

Recently uploaded

Agentic Intro and Hands-on: Build your first Coded Agent