A practical guide to building agents (OpenAI)

Agents are systems that independently accomplish tasks on your behalf. (Introduction p.4)

Conclusion (p.32)

To build reliable agents, start with strong foundations: pair capable models with well-defined tools and clear, structured instructions.

Use orchestration patterns that match your complexity level, starting with a single agent and evolving to multi-agent systems only when needed.

Start small, validate with real users, and grow capabilities over time.

When should you build an agent?

Complex decision-making:

Difficult-to-maintain rules:

Heavy reliance on unstructured data:

Agent design foundations

3つのcore components

Model

Tools

Instructions

Best practices for agent instructions

Use existing documents

Prompt agents to break down tasks

Define clear actions

Capture edge cases

You can use advanced models, like o1 or o3-mini, to automatically generate instructions from existing documents.

code:txt

You are an expert in writing instructions for an LLM agent. Convert the

following help center document into a clear set of instructions, written in

a numbered list. The document will be a policy followed by an LLM. Ensure

that there is no ambiguity, and that the instructions are written as

directions for an agent. The help center document to convert is the

following {{help_center_doc}}

Orchestration

single agent / multi agents

triage_agent (p.22-23)

Guardrails

それでも無理になったところから分離するという方針を採ったほうが良いという提言ですな。