ChatGPT: Advanced Workflows

The advanced rung: multi-step chains, working over long source material, custom assistants, and testing prompts like software.

Working over long documents and data

When the source is bigger than the conversation Long sources create two failures: material silently falls out of view, and the model answers about the parts it happens to be attending to. Both are manageable. Map, then dive Do not ask a 90-page document a detailed question straight away. First ask for a map: “List […]

When the assistant needs to look things up

The limits of an assistant that only has files A custom assistant with reference documents attached is genuinely useful, and it hits a wall the moment someone asks about something that changes. Order status, current stock, this month’s figures, whether a customer is still on the old plan — none of that lives in an […]

Costs, limits and what breaks at scale

What worked for ten items behaves differently at ten thousand An advanced workflow that runs beautifully in a chat window meets a different set of problems when it runs unattended over real volume. None of them are exotic; all of them surprise people the first time. Cost is per call, and prompts are charged every […]