⏱️ 12 min read◆📊 Intermediate-Advanced◆🎯 Creativity & Ideation
Most brainstorming is broken. Group dynamics suppress bold ideas. The first ideas anchor the conversation. Social pressure kills divergent thinking. Nobody evaluates ideas systematically — they just vote with enthusiasm. The result is a list of incremental suggestions that look a lot like what was already being done. Claude changes the brainstorming equation entirely: no social stakes, no fatigue, no political bias — and when given the right structure, consistently finds the directions most humans wouldn't explore without the pressure of seeming foolish.
Quick Answer
Creative ideation with Claude follows the diverge-then-converge pattern: generate volume without judgment first, then evaluate rigorously. The chapter covers five divergent techniques — volume-first generation, constraint-based ideation, perspective shifting, analogical thinking, and SCAMPER — three evaluation frameworks (impact-effort matrix, Six Hats, stress testing), and three refinement methods (progressive elaboration, combination, and building to a brief). Claude is not replacing creativity; it is creating the conditions for creativity to work better.
What Is the Diverge-Then-Converge Pattern for Creative Ideation?
All effective ideation follows the same two-phase pattern, and almost all failed brainstorming collapses these two phases into one.
The Ideation Pattern
Phase 1 — DIVERGE: Generate as many ideas as possible without judgment → Volume is the goal. Strangeness is welcome. Filtering is forbidden.
Phase 2 — CONVERGE: Evaluate, combine, and refine the best ideas → Rigorous criteria. Explicit scoring. Systematic selection.
Claude is particularly effective in the diverge phase because it has none of the social constraints that limit human generation. But it needs structure: the two phases must be kept explicitly separate, or the session collapses back into the broken group-dynamics pattern where evaluation kills divergence.
What Are the Five Divergent Ideation Techniques?
Technique 1: Volume-First Generation
The foundation of effective ideation: generate a large volume before judging anything. The first ten ideas are almost always obvious. Ideas eleven through twenty get more interesting. Ideas twenty-one through thirty are where creative breakthroughs often hide — because the obvious has been exhausted and genuinely different thinking is forced.
Claude Prompt
TOPIC: [Your challenge or opportunity]
CONTEXT:
[Brief description of the situation, constraints, and what you're
trying to achieve]
I want to generate a large volume of ideas before evaluating.
Please generate 30 ideas for [topic].
Rules:
- No filtering — include obvious AND unusual ideas
- No repetition — each idea should be genuinely distinct
- Range from completely safe to completely wild
- Include ideas that seem impractical — those often lead to
the best refinements
- Label each idea 1-30 for easy reference
After generating 30, identify:
- The 5 most conventional ideas (safe but reliable)
- The 5 most unconventional ideas (unusual but interesting)
- The 5 that surprised you most while generating
Technique 2: Constraint-Based Ideation
Paradoxically, constraints produce more creative ideas than total freedom. When all limits are removed, people default to safe, familiar solutions. Constraints force creative problem-solving by eliminating the obvious paths. The inversion round — generating the worst possible version — is especially powerful: understanding how to create the worst version reveals the essential ingredients of the best.
Claude Prompt
Generate ideas for [challenge], but with EXTREME CONSTRAINTS:
Round 1 — The Impossible Constraint:
What if it had to cost absolutely nothing?
Generate 10 ideas that work with zero budget.
Round 2 — The Speed Constraint:
What if it had to be implemented in 48 hours?
Generate 10 ideas that could launch this week.
Round 3 — The Scale Constraint:
What if it had to work for 10 million users from day one?
Generate 10 ideas that scale without changing fundamentally.
Round 4 — The Inverse Constraint:
What if we tried to achieve the OPPOSITE of our goal?
Generate 10 ideas for the worst possible version.
(Inversion often reveals what matters most by eliminating it)
After all four rounds:
What ideas appear in multiple rounds?
Which constraint produced the most interesting thinking?
Are there ideas that could be combined across rounds?
Technique 3: Perspective Shifting
Different perspectives unlock ideas that a single viewpoint never accesses. A customer with no category experience, a company from a completely different industry, someone looking back from ten years in the future, a resource-constrained startup, and the most demanding possible customer all see different things — and the gap between those views is where novel ideas live.
Claude Prompt
Generate ideas for [challenge] from these perspectives:
PERSPECTIVE 1: The Customer Who Has Never Used This Category
Someone who has zero experience with this type of product/service.
What would they instinctively want? What would confuse them?
What would feel obviously missing? (5 ideas)
PERSPECTIVE 2: A Company in a Completely Different Industry
How would [Amazon / Apple / a hospital / a hotel chain] approach this?
What would they do differently based on their core competency? (5 ideas)
PERSPECTIVE 3: Someone in 2035
Looking back from 10 years in the future, what seems obvious now
that we're missing? What would they consider primitive about our
current approach? (5 ideas)
PERSPECTIVE 4: A Resource-Constrained Startup
With 3 people and $50K, how would they solve this problem?
What would they ruthlessly prioritize? What would they skip entirely? (5 ideas)
PERSPECTIVE 5: The Most Demanding Customer
Someone who would reject 95% of solutions as not good enough.
What would satisfy them? What would they demand that nobody
currently provides? (5 ideas)
After all perspectives:
Which perspective generated the most useful thinking?
What insights only emerged when you took that perspective?
Technique 4: Analogical Thinking
Great ideas often come from importing solutions from unrelated domains. The question isn't "what's a good idea?" but "who has already solved something structurally similar — and what was their insight?" Nature, different industries, historical examples, and unexpected metaphors each provide a different kind of transfer.
Claude Prompt
I'm trying to solve: [Your challenge]
Find analogies in unexpected domains:
NATURE:
How does nature solve this type of problem?
(Competition for resources / navigation / adaptation / symbiosis)
What biological system maps to our challenge?
Translate 3 natural solutions into our context.
DIFFERENT INDUSTRIES:
Which industry faces a structurally similar challenge?
How do they solve it?
What can we steal and adapt? (Give 3 examples)
HISTORICAL EXAMPLES:
Has this type of problem been solved before in history?
What was the breakthrough insight then?
Does that insight apply here?
UNEXPECTED METAPHORS:
If our challenge were a game, what kind of game is it?
If our challenge were a weather pattern, what would it be?
If our challenge were a physical object, what would it look like?
Sometimes the most useful ideas come from the metaphor that makes
you say "that's exactly right" before you understand why.
Technique 5: SCAMPER
SCAMPER generates ideas by systematically applying seven transformations to an existing concept. Each letter prompts a genuinely different kind of thinking, and combinations of transformations often produce the most novel results.
S
Substitute
Replace a component with something else
C
Combine
Merge with something unexpected
A
Adapt
Adjust to fit a new context
M
Modify / Magnify
Change a quality — bigger, faster, louder
P
Put to Other Uses
Find new applications or audiences
E
Eliminate
Remove a component entirely
R
Reverse / Rearrange
Flip the process or relationship
Claude Prompt
Apply the SCAMPER technique to [existing product/service/process]:
SUBSTITUTE:
What components could be replaced with something else?
What materials, processes, people, or places could be substituted?
(Generate 3 ideas)
COMBINE:
What could be merged with something else?
What would happen if you combined this with [something unexpected]?
(Generate 3 ideas)
ADAPT:
What could be adjusted to fit a new context?
What ideas from elsewhere could be adapted here?
(Generate 3 ideas)
MODIFY / MAGNIFY / MINIMIZE:
What happens if you change a quality — make it larger, smaller,
faster, slower, louder, quieter?
What if you exaggerated one feature to an extreme?
(Generate 3 ideas)
PUT TO OTHER USES:
What else could this be used for?
Who else could use this in a way we haven't considered?
(Generate 3 ideas)
ELIMINATE:
What happens if you remove a component entirely?
What's the simplest possible version?
(Generate 3 ideas)
REVERSE / REARRANGE:
What if you reversed the process or the relationship?
What if the customer became the provider?
What if the last step became the first?
(Generate 3 ideas)
After SCAMPER: Which transformation produced the most interesting ideas?
Are there combinations of transformations that create something genuinely new?
How Do You Evaluate Ideas Rigorously?
Generating ideas is only half the work. The other half is evaluating them rigorously — separating interesting-sounding ideas from genuinely valuable ones. Three frameworks each catch different types of problems.
Evaluation Framework 1: The Impact-Effort Matrix
Claude Prompt
Here are the ideas we generated: [paste your ideas list]
Evaluate each idea on two dimensions:
IMPACT (1-10): If this idea worked, how significant would the outcome be?
Consider: size of problem solved, value created, scale of benefit
EFFORT (1-10): How much work, cost, and risk would execution require?
Consider: resources needed, complexity, time to implement
(1 = very easy, 10 = extremely difficult)
Plot each idea in one of four quadrants:
QUICK WINS (High Impact, Low Effort):
These ideas should be done immediately. List them with rationale.
BIG BETS (High Impact, High Effort):
These are strategic priorities. Evaluate carefully before committing.
FILL-INS (Low Impact, Low Effort):
Do these only when resources are available and priorities are covered.
MONEY PITS (Low Impact, High Effort):
These should be eliminated or fundamentally rethought.
After mapping:
What patterns emerge? Which ideas are clustered together?
Are there Quick Wins we're overlooking?
Which Big Bets are worth the investment?
Evaluation Framework 2: The Six Thinking Hats
Edward de Bono's Six Thinking Hats evaluates ideas by deliberately taking six different mental perspectives in sequence. Each hat catches what the others miss — which is why the Black Hat (critical judgment) must follow the Yellow Hat (optimistic value), not precede it.
⬜ White Hat — Facts
What data and information is available? What's missing?
🔴 Red Hat — Emotions
What's the gut reaction? What emotions does it trigger?
⬛ Black Hat — Caution
What could go wrong? Flaws, risks, and reasons to fail.
🟡 Yellow Hat — Value
What's the best case? The strongest possible argument for it.
🟢 Green Hat — Creativity
How could this be improved? What variations are possible?
🔵 Blue Hat — Synthesis
Balanced conclusion from all perspectives. Recommended next step.
Claude Prompt
Evaluate this idea using the Six Thinking Hats:
IDEA: [Describe the idea clearly]
CONTEXT: [What problem it solves, who it serves]
WHITE HAT (Facts):
What data and information do we have about this idea?
What information is missing that we need?
What do the facts alone tell us?
RED HAT (Emotions/Intuition):
What's the gut reaction to this idea?
What emotions does it trigger in users, stakeholders, team?
What's the initial enthusiasm level?
BLACK HAT (Critical Judgment):
What could go wrong?
What are the flaws, risks, and problems?
Why might this fail? (Be harsh — this is the most valuable hat)
YELLOW HAT (Optimistic Value):
What's the best case scenario?
What value does this create if it works?
What's the strongest possible case FOR this idea?
GREEN HAT (Creative Possibilities):
How could this idea be improved?
What variations are possible?
What creative combinations could make this stronger?
BLUE HAT (Process/Synthesis):
What is the balanced conclusion from all perspectives?
What's the recommended next step?
What additional thinking is needed before deciding?
After all six hats:
Is this idea worth pursuing? What would make it better?
Evaluation Framework 3: Idea Stress Testing
Before committing resources to an idea, stress-test it against the hardest conditions it will face. The goal isn't to kill the idea — it's to surface every weakness before development begins, when changes are cheapest.
Claude Prompt
Stress-test this idea: [Idea description]
ASSUMPTION TEST:
What 5 assumptions must be true for this idea to work?
Rate each assumption: High / Medium / Low confidence
Which assumption being wrong would kill the idea?
DEVIL'S ADVOCATE:
Make the strongest possible argument AGAINST this idea.
What would a skeptic say?
What would a direct competitor do to counter it?
EDGE CASE TEST:
How does this idea perform for:
- The most demanding user
- The least sophisticated user
- At 10x the expected scale
- Under resource constraints (half the budget/time)
- When a key assumption changes
TIMING TEST:
Why now? What makes this the right moment?
What if we waited 12 months — would the idea be better or worse?
What if we had to do it in 30 days — what would have to be true?
UNIQUENESS TEST:
Has this been tried before?
If yes: why did it fail? What's different about our approach?
If no: why hasn't it been tried? Is that a gap or a warning sign?
VERDICT:
Given this stress test — pursue, modify, or abandon?
If modify: what changes would make it significantly stronger?
Key Insight
The Uniqueness Test question "why hasn't it been tried?" is particularly important. The two possible answers — "nobody saw this gap" and "everyone tried and failed" — require completely different responses. The former is an opportunity; the latter demands understanding the reason for prior failure before proceeding.
How Do You Refine and Develop Raw Ideas?
A good idea at conception is rarely a great idea at execution. Refinement is where good ideas become great ones — and where most teams lose momentum by either developing ideas too early (before evaluating all options) or never developing them at all (leaving them as loose concepts on a list).
Refinement Technique 1: Progressive Elaboration
Develop an idea in four layers — each layer adds complexity only after the simpler version is solid. This prevents over-engineering early and ensures the core insight remains clear throughout development.
Claude Prompt
We have this idea: [Describe the core idea]
Develop it progressively:
LEVEL 1 — Core Concept (50 words):
What is the single essential insight of this idea?
Strip everything away — what remains?
LEVEL 2 — Basic Version (200 words):
What's the simplest version of this idea that still delivers value?
What's the minimum viable concept?
LEVEL 3 — Full Version (500 words):
What does the complete, well-executed version look like?
What features, components, or elements make it whole?
LEVEL 4 — Exceptional Version:
What would make this truly outstanding vs just good?
What's the detail or element that elevates it from 7/10 to 10/10?
At each level: what does the user experience feel like?
What problem are they solving at each level of elaboration?
Refinement Technique 2: The Combination Method
Often the best ideas aren't new concepts — they're unexpected combinations of existing ones. Systematic combination explores pairings that intuition would never naturally generate.
Claude Prompt
Here are our top ideas from the brainstorm: [List 10 ideas]
Systematically combine pairs:
For each combination below, explore what a merged version would look like:
- Idea 1 + Idea 4
- Idea 2 + Idea 7
- Idea 3 + Idea 9
- [You choose 2-3 more combinations that seem interesting]
For each combination:
- What's the core insight that emerges from merging?
- Is the combination stronger or weaker than either idea alone?
- What new problem does the combination solve?
- What's the simplest way to execute the combination?
After exploring all combinations:
Which combination produces something genuinely new?
Which ideas are actually better combined than standalone?
Refinement Technique 3: Building to a Brief
When an idea needs to be developed for a specific purpose, a structured brief forces every dimension into the open — including the risks and the first testable step.
Claude Prompt
Develop this idea into a complete brief:
CORE IDEA: [What the idea is]
PURPOSE: [What it's for — product launch, campaign, new feature, etc.]
AUDIENCE: [Who it's for]
CONSTRAINTS: [Budget, timeline, resources]
Please build out:
1. CONCEPT STATEMENT (2 sentences):
What is this? Why does it matter?
2. PROBLEM IT SOLVES:
What pain, frustration, or gap does this address?
Why do people care about this being solved?
3. HOW IT WORKS:
Step-by-step explanation of the mechanism
What the user does, what happens, what they get
4. WHY IT'S DIFFERENT:
What's the key distinction from existing alternatives?
What's the genuine insight here?
5. WHO IT'S BEST FOR:
Specific person or use case where this excels
Who should NOT use this
6. RISKS AND MITIGATIONS:
Top 3 risks to success
How each could be addressed
7. FIRST STEP:
What's the single most important next action to test or develop this idea?
How Do You Use Claude to Prepare for and Process Team Brainstorming?
Claude is useful before a team session — to sharpen the framing and generate stimulus — and after the session, to cluster, surface hidden gems, and recommend priorities from the raw output.
Before the Session
Claude Prompt
I'm running a brainstorming session with my team on: [Topic]
Please help me prepare:
1. FRAMING THE CHALLENGE:
Help me write a "How Might We" statement that's specific enough
to focus but open enough to allow creative responses.
Bad: "How might we improve our product?"
Better: "How might we make first-time users feel successful in their
first 5 minutes, even without any prior knowledge?"
Generate 3 alternative framings of our challenge.
2. WARM-UP PROMPTS:
Generate 5 warm-up questions that get people thinking creatively
about the topic area without priming specific answers.
3. DIVERSE STIMULUS:
What 3 examples from completely different industries or contexts
could I show the team to stimulate unexpected thinking?
4. PROVOCATIVE QUESTIONS:
Generate 5 "what if" questions that challenge our assumptions:
"What if we charged customers to NOT use our product?"
"What if the customer service team made all product decisions?"
[Generate 5 more in this spirit]
After the Session
Claude Prompt
Our brainstorming session produced these ideas: [paste raw list]
Please help me process them:
1. CLUSTERING:
Group these ideas into 4-6 themes.
Name each theme with a memorable label.
Which theme has the most ideas? The most diverse ideas?
2. HIDDEN GEMS:
Which ideas seem unusual or overlooked but might have real potential?
What would need to be true for those to be valuable?
3. INTEGRATION OPPORTUNITIES:
Which ideas from different groups could be combined into something
stronger than either part?
4. PRIORITIZATION:
Based on the themes and ideas, recommend 3 ideas to pursue first.
Explain your reasoning — not just what to do but why these three.
5. NEXT STEPS:
For each of the 3 priority ideas, what's the minimum viable test?
How could we quickly validate or invalidate the core assumption?
Key Insight
The "Hidden Gems" prompt consistently surfaces ideas that got overlooked in live discussion — typically because they were quiet, odd-sounding, or came from someone with less status in the room. Claude has no social hierarchy, so it evaluates the idea on its own merits rather than its source.
What Are the Most Common Brainstorming Mistakes?
❌ Mistake 1: Evaluating While Generating
"That idea won't work because of [reason]" — said during ideation.
✅ Correct Approach
Complete the diverge phase fully before any evaluation. The two phases must be kept separate.
❌ Mistake 2: Stopping at 10 Ideas
"We have 10 ideas — that's enough."
✅ Correct Approach
The most interesting ideas often come after 20–25. Push through the obvious zone.
❌ Mistake 3: Only Asking for Good Ideas
"Give me your best ideas for X."
✅ Correct Approach
"Give me 30 ideas including wild and impractical ones." The wild ones often contain the seed of the great ones.
❌ Mistake 4: Skipping Evaluation
"We'll figure out which ideas are good later." Later never comes.
✅ Correct Approach
Always evaluate using a consistent framework immediately after generating.
❌ Mistake 5: Developing Ideas Too Early
Spending two hours developing idea #3 before evaluating all 30.
✅ Correct Approach
Quick evaluation of all ideas first, then deep development of the top three to five.
Key Takeaways
Diverge fully before converging — Separate generation from evaluation completely
Volume unlocks quality — The 21st idea is usually more interesting than the 3rd
Constraints spark creativity — Limitations force more imaginative solutions
Perspective shifting finds blind spots — What's obvious from one angle is invisible from another
Evaluation needs frameworks — Impact-effort, Six Hats, stress testing each catch different things
Refinement is where ideas become valuable — Raw ideas are starting points, not endpoints
Combination produces breakthroughs — The best ideas often merge two seemingly unrelated ones
Your Turn: Assignment
Challenge: Run a real ideation session on a challenge currently in progress.
Beginner Option
Use the Volume-First technique (30 ideas) on a challenge that's been stuck. Identify the five most unconventional ideas and pick one to develop using Progressive Elaboration.
Intermediate Option
Use the Constraint-Based technique (all four rounds) on a creative challenge. Which constraint produced the most surprising ideas? Apply the Impact-Effort matrix to evaluate the results.
Advanced Option
Run a full ideation session: diverge with three different techniques, evaluate with two different frameworks, refine the top three ideas using the Complete Brief template, then stress-test the strongest one.
Reflection Questions
At what idea number did things start getting genuinely interesting? Which constraint produced the most unexpected thinking? Which perspective — first-time customer, different industry, future-looking — revealed something previously unconsidered?
In Chapter 19: Vision & Multimodal Prompting, the series turns to Claude's vision capabilities — analyzing images, extracting information from screenshots, combining visual and text prompts, and building multimodal workflows.
Frequently Asked Questions
Most brainstorming fails because group dynamics suppress bold ideas, the first ideas anchor the conversation, social pressure kills divergent thinking, and ideas are never evaluated systematically. Claude removes social stakes entirely — it has no political interest in an idea winning or losing, can generate volume without fatigue, and can evaluate against explicit criteria without bias. The key is structuring sessions correctly: a full diverge phase before any convergence, explicit evaluation frameworks, and refinement after selection.
All effective ideation follows a two-phase pattern: diverge first by generating as many ideas as possible without judgment — volume is the goal, strangeness is welcome, filtering is forbidden — then converge second by evaluating, combining, and refining the best ideas with rigorous criteria. Most people collapse these phases, evaluating while generating, which kills divergence. Or they never converge properly, leaving ideas with no selection logic. Claude excels in the diverge phase because it generates without social constraints, but both phases need explicit structure.
When all limits are removed, people default to safe, familiar solutions. Constraints force creative problem-solving by eliminating the obvious paths and requiring genuinely different thinking. The constraint-based ideation technique uses four rounds: a zero-budget round, a 48-hour implementation round, a 10-million-user scale round, and an inversion round — generate the worst possible version. The inversion round is especially powerful: understanding how to create the worst version reveals the essential ingredients of the best version.
SCAMPER generates ideas by applying seven systematic transformations to an existing product, service, or process: Substitute (replace components with something else), Combine (merge with something unexpected), Adapt (adjust to fit a new context), Modify or Magnify or Minimize (change a quality — size, speed, loudness), Put to other uses (find new applications), Eliminate (remove a component entirely), and Reverse or Rearrange (flip the process or relationship). Each transformation generates 3 ideas, and combinations of transformations often produce the most genuinely new concepts.
The Six Thinking Hats framework evaluates ideas by deliberately taking six different mental perspectives in sequence: White Hat considers facts and missing information; Red Hat captures gut reactions and emotional responses; Black Hat makes the harshest critical case against the idea; Yellow Hat builds the strongest possible case for the idea; Green Hat explores creative improvements and variations; Blue Hat synthesizes all perspectives into a balanced conclusion and recommended next step. Each hat catches different weaknesses and strengths that a single-perspective evaluation misses.