A businessleader’s guideto workingwith agentsA businessleader's guideto workingwith agents Preparing your teams for the shift to agentsPreparing your teams for the shift to agents ContentsContents Introduction3Introduction3 What is an AI agent?5What is an Al agent?5 How models shape an agent’s abilities7How models shape an agent's abilities7 ToolsTools 99 Guardrails11Guardrails11 Workflow automation vs. agents13Workflow automation vs. agents13 Preparing your team to work with agents16Preparing your team to work with agents16 Conclusion23Conclusion23 IntroductionIntroduction A year ago, the term “AI agent” was largely abstract, used more to describe potential than practice.Today, businesses are shifting from curiosity to integration: startups and enterprises alike arebuilding agents, organizations are embedding them in workflows, and employees are alreadyexpanding the scope of what they can achieve. Ayear ago,the term"Al agent"was largely abstract, used more to describe potential than practice.Today,businesses are shifting from curiosity to integration: startups and enterprises alike arebuilding agents, organizations are embedding themin workflows,and employees are alreadyexpandingthescopeofwhattheycanachieve. Yet the term is still fuzzy. We’re regularly asked what an agent is, how to work with them, and howto plan for them.Yet the term is still fuzzy. We're regularly asked what an agent is, how to work with them, and howto plan for them. In this guide, we willIn this guide, we will 01Define AI agents and their core components02Show how they work and where they add value03Outline steps for delegating, supervising, and optimizing agents04Demonstrate how to organize agents and tips to avoid unnecessary complexity01Define Al agents and their core components02Show how they work and where they add value03Outlinestepsfordelegating,supervising,andoptimizing agents04Demonstrate howtoorganizeagentsandtipsto avoid unnecessary complexity This shift is happening quickly. When used well, agents free people to think deeply, work creatively,and solve problems together. Now is the time to experiment, understand how agents work, andcreate guardrails that ensure safe, valuable use.This shift is happening quickly.When used well, agents free people to think deeply, work creativelyand solve problems together.Now is the time to experiment, understand how agents work, andcreateguardrailsthatensure safe,valuable use. There are many types of agents. This guide focuses on using pre-built agents like ChatGPT inagent mode, but you can also create custom agents - whether for internal productivity or topower new revenue-generating features in your product. For guidance on building customagents, see our Practical Guide to Building Agents, or checkout AgentKit, our full-stack agentplatform. We’ll share more guidance on how to scope, build, and measure the success ofagents in a future guide.There are many types of agents.This guide focuses on using pre-built agents like ChatGPT inpower new revenue-generating features in your product. For guidance on building customagents,seeourPracticalGuide toBuildingAgents,or checkout AgentKit, ourfull-stack agentplatform.We'll share more guidance on howto scope, build, and measure the success ofagents inafuture guide. What is anAI agent?What is anAl agent? An AI agent is a system that can plan, decide, and act independently to achieve a goal whileoperating within guardrails set by humans. Agents connect reasoning with tools, adapt mid-taskbased on information they encounter, and can carry complex tasks end-to-end. An Al agent is a system that can plan, decide, and act independently to achieve a goal whileoperating within guardrails set by humans.Agents connect reasoning with tools, adapt mid-taskbased on informationthey encounter, and can carry complextasks end-to-end. Agents are built from three key elements:Agentsarebuiltfromthreekeyelements: Together, these give agents the capacity for autonomous decision-making and action. We expandon how agents plan, act, and adapt, and how this differs from workflows in the sections ahead.Together,these give agents the capacity for autonomous decision-making and action. We expandon how agents plan, act, and adapt,and how this differs from workflows in the sections ahead. What is an AI agent?What is an Al agent? What AI agents are not What Al agents are not Many AI systems can answer questions or automate tasks, but they don’t plan, adapt, or carrywork forward independently, that’s where agents come in. Many Al systems can answer questions or automate tasks, but they don't plan, adapt, or carryworkforward independently,that's where agents comein. For example, a Q&A chatbot can return answers and even make tool calls to gather the mostaccurate, up-to-date information—but it can’t plan or adapt. Imagine asking, “What’s our travelpolicy?” It could pull the latest version from your HR system and provide a clear answer, but itwouldn’t be able to