ChatGPT Advanced tutorial

Advanced · step 1 of 7

Decompose and chain

1. Split a big job into steps with a clear output each.
2. Feed each step's output into the next.
3. Check the output between steps, not just at the end.

Why this step matters: Small steps with a checked output each fail visibly and early, instead of one big answer that is wrong somewhere.

Advanced · step 2 of 7

Structured output

Ask for a fixed table or JSON so you can check and reuse it.

Return only JSON: [{"claim": "...", "source_url": "...", "quote": "...", "confidence": "high|medium|low"}]

A good result: only JSON, no text before or after it, one object per claim. Paste it into any JSON validator; if it doesn't parse, reply "Return valid JSON only" and run it again.

Why this step matters: A fixed format can be checked by eye or by code and reused without retyping.

Advanced · step 3 of 7

Automation and agents

Scheduled tasks are on paid plans only (OpenAI). Custom GPTs and Codex are listed on Plus. Read anything it plans to send before it goes out.

Why this step matters: Automation repeats mistakes at scale, so you need to know exactly what is run and what is checked.

Advanced · step 4 of 7

Failure modes to watch

• Confident made-up citations.
• Agreeing with a wrong premise in your question (sycophancy).
• Losing early instructions in very long chats.
• Memory saving a wrong fact that then colours later answers.

Why this step matters: Knowing how it typically fails tells you where to look when you review its work.

Advanced · step 5 of 7

Cost control

OpenAI's pricing page didn't show dollar prices as readable text when we checked, so look at checkout. Test Free for a week with real tasks before paying. The API is billed separately from ChatGPT.

Context on Free is about 27K tokens (~12 pages), per OpenAI's pricing page. In a long chat, early parts fall out, so start fresh with a summary.

Why this step matters: Paying for features you don't use is the most common waste. Knowing the limits first tells you if the free tier is enough.

Advanced · step 6 of 7

Privacy and team rules

• Don't paste passwords, ID numbers, card numbers or other people's private details.
• Assume the company can read and keep your chats unless its official policy says otherwise.
• For health, money or legal decisions, confirm with a professional or an official source.
• Write down which data is allowed in which tool, and who approves paid plans.

Why this step matters: Anything you paste may be stored and read. Deciding what never goes in is easier than undoing a leak.

Advanced · step 7 of 7

Build a 10-task test set

1. Collect 10 real tasks with known good answers.
2. Run them now and score each pass/fail.
3. Re-run them when the bot, model or your prompt changes.
4. Keep the prompt that wins, not the one that feels better.

Why this step matters: Without a fixed test set, 'better' just means 'felt better'. Scores let you compare prompts and models fairly.