Prompt Chaining: How to Break Complex Tasks Into Simple Steps (2026 Guide)

Research the topic "{{topic}}" for a {{target audience}} audience.

Find and summarize:
- 5-7 key facts, statistics, or findings (with approximate sources)
- 3 common misconceptions about this topic
- 2-3 expert perspectives or contrarian viewpoints
- What most existing content on this topic gets wrong or misses

Output as a structured research brief. No narrative — just organized findings I can reference.

Using the research brief below, create a detailed outline for a {{word count}}-word {{content type}} targeting {{audience}}.

RESEARCH BRIEF:
{{paste Step 1 output}}

The outline should include:
- A working title and 1-sentence thesis
- Hook angle for the introduction
- 5-7 main sections with H2 headings
- 2-3 sub-points under each section
- Where to place statistics, examples, and expert quotes from the research
- A specific call-to-action for the conclusion

Format as a nested bullet list I can hand to a writer.

Write a full draft based on the outline below. Target {{word count}} words.

OUTLINE:
{{paste Step 2 output}}

Writing guidelines:
- Tone: {{describe tone}}
- Open with a hook, not a throat-clearing paragraph
- Use specific examples over general claims
- Write section transitions that create forward momentum
- No filler phrases: "it's important to note," "in today's world," "at the end of the day"
- End with a concrete next step, not a summary

Write the complete draft. Do not skip sections or use placeholders.

Edit the draft below for clarity, flow, and impact. Do not rewrite — improve.

DRAFT:
{{paste Step 3 output}}

Editing checklist:
- Cut every sentence that doesn't earn its place
- Replace vague claims with specific ones
- Fix any awkward transitions between sections
- Ensure the opening hook works in the first 2 sentences
- Verify the conclusion delivers a clear next step
- Flag any claims that need a source citation

Return the edited draft with your changes tracked in bold. Add a brief editor's note at the end listing the 3 biggest changes you made and why.

I need to build {{feature description}} for a {{tech stack}} application.

Analyze these requirements and produce:
1. A list of functional requirements (what the feature must do)
2. A list of non-functional requirements (performance, security, edge cases)
3. Questions or ambiguities I should resolve before coding
4. Input/output specifications for the main function(s)

Be specific. If I said "user authentication," I want to know: OAuth or email/password? Session or JWT? What happens on failed login?

Based on the requirements below, design the technical implementation.

REQUIREMENTS:
{{paste Step 1 output}}

Produce:
- File structure (which files to create or modify)
- Key data models / types / interfaces
- API endpoints or function signatures with parameters and return types
- Flow diagram in text (step 1 → step 2 → etc.)
- Edge cases and how each is handled
- Dependencies or libraries needed

Do not write code yet. This is the blueprint.

Implement the design below in {{language/framework}}.

DESIGN:
{{paste Step 2 output}}

Write complete, production-ready code. Not pseudocode. Not stubs. Every function implemented, every edge case handled, every type defined.

Follow these conventions:
- {{coding style notes — e.g., use TypeScript strict mode, prefer async/await, use named exports}}
- Include error handling for every external call
- Add JSDoc/docstring comments on public functions
- Use descriptive variable names

Output each file separately with its full path as a header.

Write comprehensive tests for the code below using {{test framework}}.

CODE:
{{paste Step 3 output}}

Cover:
- Happy path for each public function
- Edge cases identified in the design
- Error handling paths (invalid input, network failures, auth errors)
- Boundary conditions (empty arrays, null values, max limits)

Each test should have a descriptive name that explains what it verifies. Group tests by function or feature.

I have a dataset with the following structure:

{{describe columns, data types, row count, and source}}

Sample rows:
{{paste 5-10 sample rows}}

Analyze this dataset and tell me:
1. What each column likely represents
2. Potential data quality issues (nulls, outliers, inconsistencies)
3. The 5 most interesting questions this data could answer
4. Which columns are most likely correlated
5. Suggested cleaning steps before analysis

Based on the data profile below, perform the following analyses:

DATA PROFILE:
{{paste Step 1 output}}

Analyses to run:
1. Descriptive statistics for all numeric columns
2. Distribution analysis for the top 3 variables
3. Correlation analysis between {{variable A}} and {{variable B}}
4. Trend analysis over {{time period}} if time data exists
5. Segment comparison: {{group A}} vs {{group B}}

For each analysis, state:
- What you found
- Whether the finding is statistically meaningful or just noise
- One business implication

Use plain language. The audience is {{business role, e.g., "a marketing director"}}, not a data scientist.

Based on the analysis below, design a visualization dashboard.

ANALYSIS:
{{paste Step 2 output}}

For each key finding, recommend:
- Chart type and why it's the best choice
- X axis, Y axis, color encoding, and any filters
- Title that communicates the insight (not just the data)
- One callout annotation per chart highlighting the key takeaway

Output as a specification I can hand to a designer or implement in {{tool — e.g., Tableau, Python matplotlib, D3.js}}.

Write an executive summary report based on the analysis and visualization spec below.

ANALYSIS: {{paste Step 2 output}}
VISUALIZATION SPEC: {{paste Step 3 output}}

Report structure:
- One-paragraph executive summary (the single most important finding)
- 3-5 key insights with supporting data
- Recommended actions for each insight
- Risks or caveats the reader should know
- Suggested next analysis

Write for {{audience role}}. Maximum 800 words. Every sentence should either present a finding or recommend an action.

Analyze this job posting and extract the core competencies:

{{paste job posting}}

Produce:
1. The 5 most critical skills for this role (ranked)
2. The 3 non-obvious qualities that will separate good from great
3. Red flags to watch for in candidates
4. What this role will actually spend 80% of their time doing (vs. what the posting says)

Based on the role analysis below, create a structured interview question set.

ROLE ANALYSIS:
{{paste Step 1 output}}

Generate:
- 3 behavioral questions per critical skill (using STAR format prompts)
- 2 situational questions testing the non-obvious qualities
- 1 technical/practical question that reveals real ability (not trivia)
- 1 "tell me about a failure" question specific to this domain

For each question, include:
- The question itself
- What a strong answer includes
- What a weak answer sounds like
- A follow-up probe question

Build a scoring rubric for the interview questions below.

QUESTIONS:
{{paste Step 2 output}}

For each question, create a 1-5 scale:
- 1 (Poor): Specific description of what this looks like
- 3 (Acceptable): Specific description
- 5 (Exceptional): Specific description

The rubric should be usable by any interviewer, not just the hiring manager. Avoid subjective language like "good communication" — describe the observable behavior.

Our business problem: {{describe the problem in 2-3 sentences}}

Context:
- Product: {{what your product does}}
- Current metrics: {{relevant numbers}}
- What we've tried: {{previous attempts}}

Produce:
1. A precise problem statement (one sentence)
2. Who this problem affects most and why
3. The root cause vs. the symptoms we're seeing
4. What success looks like in measurable terms

Based on the problem framing below, generate testable hypotheses.

PROBLEM FRAME:
{{paste Step 1 output}}

For each hypothesis:
- State it as "If we [action], then [measurable outcome] because [mechanism]"
- Rate confidence (high/medium/low) and effort to test
- Identify the biggest assumption embedded in this hypothesis
- Suggest the fastest way to disprove it

Generate 5-7 hypotheses. Rank by expected impact × confidence.

Design a rigorous experiment to test this hypothesis:

{{paste selected hypothesis from Step 2}}

Include:
- Test group vs. control group criteria
- Sample size needed for statistical significance (state your assumptions)
- Duration of test
- Primary metric and how to measure it
- Guardrail metrics (what must NOT get worse)
- How to handle confounding variables
- Decision criteria: what result means "ship it" vs "kill it" vs "iterate"

Based on the analysis from Step 2:

[paste Step 2 output here]

Now generate...