Learn AI · 6 lessons · Free, live now
Agentic Processes
Newer AI features do not just answer, they work: research for ten minutes, browse websites, run a series of steps. Six lessons on what is actually happening while you wait, how to brief these features well, and how to check their work before you trust it.
When you finish, you can
Run a Deep Research or agent-mode task with confidence, know what it is doing while you wait, and review the result before you trust it.
The lessons
- 01Chat versus a processA chat answers once. A process runs a loop of plan, act, check. Why that takes minutes and what the wait buys you.
- 02Deep Research explainedWhat ChatGPT, Gemini, and Claude actually do during a research run, where to find it as of August 2026, and how to write a brief.
- 03Agent mode and browsingWhen the AI opens websites and clicks for you. Where each app stands as of August 2026, and the moments it must stop and ask.
- 04Multi-step workflowsGather, draft, check, deliver. Breaking a job into steps, handing off between chats, and when one prompt is enough.
- 05Checking the workCitations that do not exist, sampling three claims, and the five-minute review that makes a research report safe to act on.
- 06Cost and timeWhat research and agent features cost as of August 2026, how long they run, and a decision table for when plain chat is better.