A sub-agent is a specialist AI agent called by a parent agent to handle one specific part of a larger task. It focuses on a single job — like writing an email or analyzing a transcript — then returns its result to the parent agent that called it.
The Specialist on the Team
Think of a small business where the owner delegates tasks. The owner doesn’t write the newsletter, design the graphics, and post on social media personally. Instead, they call on specialists: “You handle the writing, you handle the design, you handle the posting.” Each specialist does their part and reports back. In AI terms, the owner is the parent agent (or orchestrator) and each specialist is a sub-agent.
A sub-agent is intentionally limited in scope. It doesn’t need to understand the full workflow. It just needs to do its one job well and return the result. A transcript analysis sub-agent doesn’t need to know that the output will eventually become a blog post, an email, and five social media posts. It just analyzes the transcript and hands back its findings.
Why Sub-Agents Exist
Complex tasks work better when broken into focused pieces. A single agent trying to do everything — transcribe, analyze, write, email, publish, schedule — has to hold too much context and juggle too many tools at once. That leads to confusion and errors. Sub-agents solve this by keeping each step simple and focused.
There’s also a quality benefit. A sub-agent with specific instructions for email writing will produce better emails than a general-purpose agent that also handles ten other tasks. The narrower the focus, the better the output — same principle that makes specialist employees more effective than generalists for specific tasks.
What This Means for Educators
As a course creator or coach, sub-agents are working behind the scenes every time you run a multi-step workflow. When you trigger a content repurposing cascade, sub-agents handle each step — one extracts the transcript, another writes the summary, another drafts the email, another creates social posts. You see the finished outputs; you don’t need to manage each sub-agent individually.
If you build your own agent skills, sub-agents let you create simple, reusable building blocks. An email-writing sub-agent can be called by your course launch workflow, your weekly newsletter workflow, and your student follow-up workflow — same skill, different contexts.
The Simple Rule
A sub-agent is a specialist called by a larger agent to handle one piece of a bigger job. It does its task, returns the result, and lets the parent agent decide what happens next. Think of it as delegation within your AI team — the same principle that makes human teams effective.