What makes a task "complex"
Asking for a sandwich is easy. Asking for a whole party menu, shopping list and cooking schedule in one breath is hard. A big job has many parts, so we find the parts first.
- A task is complex when it has several parts, needs different skills (reading, deciding, writing), uses long inputs, or must follow a strict format.
- Complex tasks fail in typical ways: one part is skipped, parts get mixed up, the format drifts, or a small early error spreads into later parts.
- Anthropic's guide says Claude responds well to clear, explicit instructions and that it is useful to think of it as a "brilliant but new employee" (course 1.2).
- Anthropic's guide also says that with adaptive thinking, Claude handles most multistep reasoning internally. Explicit prompt chaining (breaking a task into sequential steps) is still useful when you need to inspect intermediate outputs or enforce a specific pipeline.
- So the first skill is decomposition: write the job as inputs, steps and outputs.
| Part | Question to ask | Example: "Summarize 20 customer emails and plan replies" |
|---|---|---|
| Input | What goes in? | 20 emails |
| Steps | What are the separate actions? | Sort by topic, find urgent ones, draft replies |
| Output | What must come out, in what format? | Table with email, topic, urgency, draft reply |
| Check | How will I know it is right? | Every email has a row; urgent ones are marked |
- Choose a real job that has many parts (a report, a plan, a batch of messages).
- Write the input in one line: what you will paste in.
- List the separate actions as verbs, one per line (sort, find, draft, check).
- Write the output format: table, list, email, or fields.
- Write one line for how you will check the result.
- Mark any step where you would want to read the result before the next step. These are your chain breaks.
- Check: someone else could follow your list without asking you questions.
Videos by other creators on YouTube, for reference. They are not made by or affiliated with The Future Market.
Read the video script
[Host] "A big prompt is usually three small jobs wearing a coat."
[Screen: one huge prompt, then split into sort, draft and check] "Here is one giant request. Let me pull it apart: sort, draft, check."
[Text on screen: Input · Steps · Output · Check] "For every big job I write four things: input, steps, output, check."
[Host] "Once the parts are clear, the prompts nearly write themselves."
Take one complex job from your week. Fill in the four-row table (input, steps, output, check) and mark two chain breaks.
ClaudeStuck? Copy a ready-made prompt for Claude
Copy this, open Claude, paste it, and Claude will walk you through the lesson step by step.
For a grown-up: Claude is for adults (18+). Copy this, paste it into Claude, and use the answer to explain the lesson to your child.
I'm learning "Prompting for Complex Tasks: chains, roles and structured outputs" on The Future Market. Lesson 1: "What makes a task "complex"". What the lesson teaches: - A task is complex when it has several parts, needs different skills (reading, deciding, writing), uses long inputs, or must follow a strict format. - Complex tasks fail in typical ways: one part is skipped, parts get mixed up, the format drifts, or a small early error spreads into later parts. - Anthropic's guide says Claude responds well to clear, explicit instructions and that it is useful to think of it as a "brilliant but new employee" (course 1.2). - Anthropic's guide also says that with adaptive thinking, Claude handles most multistep reasoning internally. Explicit prompt chaining (breaking a task into sequential steps) is still useful when you need to inspect intermediate outputs or enforce a specific pipeline. - So the first skill is decomposition: write the job as inputs, steps and outputs. - Part - Question to ask - Example: "Summarize 20 customer emails and plan replies" - Input - What goes in? - 20 emails The exercise I'm working on: Take one complex job from your week. Fill in the four-row table (input, steps, output, check) and mark two chain breaks. I'm stuck. Please: 1. Explain the main idea in plain, simple words with one everyday example. 2. Walk me through the exercise one small step at a time. Ask me what I've done so far before giving the next step. 3. Don't just give me the final answer. Help me do it myself, and check my work when I share it.
I'm a parent helping my child (about 8 years old) with a lesson called "What makes a task "complex"" from the course "Prompting for Complex Tasks: chains, roles and structured outputs". Here's what the lesson says: Asking for a sandwich is easy. Asking for a whole party menu, shopping list and cooking schedule in one breath is hard. A big job has many parts, so we find the parts first. The activity: Take one complex job from your week. Fill in the four-row table (input, steps, output, check) and mark two chain breaks. Please explain this so an 8-year-old understands: short sentences, a fun everyday example, and no hard words. Then give me 3 simple questions I can ask my child to check they understood, and a way to do the activity together, one tiny step at a time.
Sources
- Anthropic, Prompting best practices (Be clear and direct; Chain complex prompts): platform.claude.com/docs/en/build-with-claude/prompt-engineering/claud
Prompt chains
This lesson is part of full access: every course and every level for one payment.
Roles and labelled parts for long prompts
This lesson is part of full access: every course and every level for one payment.
Structured outputs (tables and JSON)
This lesson is part of full access: every course and every level for one payment.
Examples and self-checks
This lesson is part of full access: every course and every level for one payment.
Debug a failing prompt
This lesson is part of full access: every course and every level for one payment.
Put it all together
This lesson is part of full access: every course and every level for one payment.
Final project
This lesson is part of full access: every course and every level for one payment.
Quiz
Name the four parts of a decomposed task.
Input, steps, output, check.
When is prompt chaining still useful?
When you need to inspect intermediate outputs or enforce a specific pipeline.
What is the most common chaining pattern in Anthropic's guide?
Self-correction: draft, review against criteria, refine.
Why put each hand-off in a tag such as <draft>?
So the AI can tell the material from your instructions.
What does setting a role do?
It focuses the AI's behaviour and tone; it does not give it new knowledge.
What is JSON?
A plain-text format of field names and values that programs and spreadsheets can read.
Does a valid JSON shape mean the facts are right?
No. Shape and facts are checked separately.
How many examples does Anthropic's guide recommend?
3 to 5.
Why include an edge-case example?
It shows the AI what to do with tricky or missing information.
What self-check line does the guide suggest?
"Before you finish, verify your answer against [criteria]."
Format changes every time. What is a good fix?
Add examples and a fixed field list.
Why change only one thing when debugging?
So you know which change helped.
Final exam
This lesson is part of full access: every course and every level for one payment.
Certificate
This lesson is part of full access: every course and every level for one payment.