Creative Prompts for Brainstorming and Innovation
AI brainstorming prompts organized as a diverge, converge and select workflow, with evaluation criteria, an illustrative worked example, and prompts for creative blocks and product ideation.
An AI assistant can be a useful brainstorming partner. The right prompts can help you break through creative blocks, generate novel ideas, and explore possibilities you wouldn't have considered. For complete content creation workflows, see our writing prompts for content creators.
Why AI for Brainstorming?
AI's Creative Strengths
- No judgment: Won't dismiss "weird" ideas
- Rapid quantity: Can generate dozens of ideas in seconds
- Cross-domain: Draws unexpected connections
- Always available: Ready when inspiration strikes
- Tireless: Keeps going when you're stuck
AI's Creative Weaknesses
- Gravitates to the obvious. Without constraints, the first ideas tend to be the most common answers to the question.
- Repeats itself. Long lists often contain the same idea in different words.
- Doesn't know what already exists. An "original" idea may already be a product, a campaign or someone's trademark.
- Doesn't check feasibility. Cost, regulation and your team's capacity are invisible to it unless you describe them.
The workflow below is designed around those weaknesses: generate widely, then narrow with explicit criteria, then test.
A Three-Stage Workflow: Diverge, Converge, Select
Mixing generation and judgment in the same step tends to kill unusual ideas early. Keep the stages separate, ideally in separate prompts.
| Stage | Goal | What the AI does | What you do |
|---|---|---|---|
| 1. Diverge | Quantity and range | Generate many varied options under constraints | Set the brief and constraints; add your own ideas |
| 2. Converge | Structure | Remove duplicates, cluster, score against your criteria | Choose the criteria and weights; challenge the scores |
| 3. Select | Commit to a test | Design a cheap experiment for the top options | Decide what to test and own the result |
Evaluation criteria
Agree on criteria before you see the ideas, so you don't unconsciously tune them to favor one you already like. A set that works for many situations, each scored 1-3:
- Fit: how directly does it serve the stated goal?
- Feasibility: can we do it with the budget, people and time we actually have?
- Distinctiveness: is it meaningfully different from what we (and obvious competitors) already do?
- Testability: can we learn whether it works quickly and cheaply?
- Risk: what's the downside if it goes wrong (reputation, cost, legal)? (3 = low risk)
Worked example: a fictional bookshop
Illustrative example. The bookshop is fictional and the idea lists below were written for this article to show the shape of each stage. They are not recorded AI responses or measured results.
Brief: A fictional independent bookshop is quiet on weekday evenings and wants an event series that brings in people aged 20-35. Budget: about $300 per month. Two staff on evening shifts. A back room that seats 25.
Stage 1 prompt (diverge):
Generate 25 ideas for a weekday-evening event series at an independent
bookshop that wants to attract people aged 20-35.
Constraints: about $300/month, two staff, a back room seating 25.
Requirements:
- At least 8 ideas must not be author talks or book clubs.
- At least 5 should involve something attendees make or do.
- No two ideas may share the same core activity.
- One line per idea. Don't evaluate them yet.
Illustrative excerpt of Stage 1 output:
1. Silent reading hour with a playlist and tea; phones in a basket
2. "Blind date with a book" swap night: bring a wrapped book, leave with another
3. Zine-making workshop using old magazines from the shop's recycling
4. First-chapter read-aloud night: attendees read the opening page of a favorite
5. Local history walk that starts and ends at the shop
...
Stage 2 prompt (converge):
Here are the 25 ideas (plus 4 of my own). First merge any duplicates.
Then group them into 4-6 clusters. Then score each remaining idea from
1-3 on: fit, feasibility, distinctiveness, testability, risk (3 = low
risk), using only the constraints in the brief. Show the scores in a
table and flag any score you're unsure about and why.
Illustrative excerpt of Stage 2 output:
| Idea | Fit | Feasibility | Distinctiveness | Testability | Risk | Total |
|---|---|---|---|---|---|---|
| Silent reading hour | 3 | 3 | 2 | 3 | 3 | 14 |
| Book swap night | 3 | 3 | 2 | 3 | 2 | 13 |
| Zine-making workshop | 2 | 2 | 3 | 2 | 3 | 12 |
| Local history walk | 1 | 2 | 3 | 2 | 2 | 10 |
The useful part of this stage is often the flags rather than the totals. A good response notes things like: "Book swap risk scored 2 because swapped books may be in poor condition; you'd need a simple quality rule." The staff should then adjust scores where they know better. They know, for example, whether the regulars would actually come to a silent reading hour.
Stage 3 prompt (select):
We've chosen the silent reading hour and the book swap night. For each,
design the cheapest 4-week test: what we'd run, what we'd count, what
result would make us continue, and what would make us stop.
At this point the decision is the bookshop's. The AI's test design is a draft. Whether "12 attendees per session" is a success depends on the shop's own numbers.
What makes this different from asking for "great ideas"
- Constraints in the diverge prompt force range ("8 ideas must not be author talks").
- Criteria are fixed before judging, and scored in a separate step.
- Humans adjust scores with local knowledge the model doesn't have.
- The output is a test, not a verdict.
Brainstorming Prompt Frameworks
The "What If" Engine
Generate 20 "What if..." scenarios for [TOPIC/PROBLEM].
Context: [CURRENT SITUATION]
Range of ideas:
- 5 conservative (small changes)
- 10 moderate (significant changes)
- 5 wild (seemingly impossible)
Don't evaluate—just generate ideas freely.
Perspective Shift
I'm stuck on [PROBLEM]. Help me see it from new angles.
View this problem as if you were:
1. A 10-year-old child
2. An alien seeing humans for the first time
3. Someone from 100 years ago
4. Someone from 100 years in the future
5. A competitor trying to solve the same problem
6. A customer who's frustrated
What would each perspective suggest?
Constraint Flipping
I'm working on [IDEA/PROJECT].
Current constraints I'm accepting:
- [Constraint 1]
- [Constraint 2]
- [Constraint 3]
For each constraint:
1. What if it didn't exist?
2. What if we made it 10x more extreme?
3. What if we made it the core feature instead of a limitation?
Generate ideas based on each flipped constraint.
The Combination Generator
Help me combine unexpected elements:
Domain A: [YOUR DOMAIN]
Domain B: [UNRELATED DOMAIN - e.g., gaming, hospitality, nature]
Generate 10 ideas that combine concepts from both domains.
Make connections that wouldn't be obvious.
Example format:
"What if [concept from A] worked like [concept from B]?"
Ideation Session Prompts
Opening a Session
We're brainstorming solutions for: [CHALLENGE]
Context:
- Current situation: [DESCRIBE]
- Failed approaches: [WHAT WE'VE TRIED]
- Stakeholders: [WHO CARES ABOUT THIS]
- Constraints: [REAL CONSTRAINTS TO RESPECT]
Start by generating 15 rapid-fire ideas with no filtering.
Include some that feel unrealistic—we'll evaluate later.
Excavating Deeper
This initial idea is interesting: [IDEA]
Help me explore it deeper:
1. What's the core insight here?
2. How could we implement this in 3 different ways?
3. What would a minimal version look like?
4. What would an ambitious version look like?
5. Who else has done something similar we could learn from?
6. What's the first step to test this?
Combining and Evolving Ideas
Here are our top 5 ideas from brainstorming:
1. [Idea 1]
2. [Idea 2]
3. [Idea 3]
4. [Idea 4]
5. [Idea 5]
Now:
1. Find 3 ways to combine ideas from the list
2. For each idea, suggest 2 variations that strengthen it
3. Identify elements that appear in multiple ideas (these might be key)
4. Which idea has the most potential, and why?
Closing a Session
Here are the ideas we're considering taking forward:
[LIST OF IDEAS]
For each, create a brief summary:
- One sentence description
- Key strengths
- Key risks/challenges
- Who should champion this
- Next step to move forward
Format as a decision document we can share with the team.
Browse our creative prompts collection for more ideation templates.
Overcoming Creative Blocks
When You're Stuck
I'm feeling stuck on [CREATIVE PROJECT].
Current state: [WHAT I HAVE]
Where I'm blocked: [THE STICKING POINT]
What I've tried: [APPROACHES THAT DIDN'T WORK]
Help me get unstuck by:
1. Reframing the problem differently
2. Suggesting 5 unexpected starting points
3. Recommending a constraint to add (sometimes limits help)
4. Proposing a completely different approach
Fresh Perspectives
I've been working on [PROJECT] too long and can't see it clearly anymore.
Act as if you're seeing this for the first time:
[DESCRIBE PROJECT]
Tell me:
1. What stands out immediately (good or bad)?
2. What's confusing or unclear?
3. What's missing that you expected?
4. What's the strongest element to build on?
5. If you had to change one thing, what would it be?
The Opposite Game
I'm considering this approach:
[YOUR APPROACH]
Generate the opposite approach. Then identify:
1. What does the opposite make you realize about the original?
2. Is there something valuable in the opposite?
3. Could you combine elements of both?
Innovation Prompts
Problem Reframing
We're trying to solve: [PROBLEM STATEMENT]
But are we solving the right problem? Help me question assumptions:
1. Rephrase this problem 5 different ways
2. What's the problem behind the problem?
3. Who defined this as the problem? Could they be wrong?
4. If this problem magically disappeared, what would the new world look like?
5. What if this isn't a problem but an opportunity?
First Principles Thinking
Help me think about [TOPIC] from first principles.
Current approach: [HOW IT'S USUALLY DONE]
1. What are the fundamental truths we can't argue with?
2. Strip away conventions—what remains?
3. If we started from scratch today, what would we build?
4. What would [Elon Musk / Jeff Bezos / nature] do?
Trend Synthesis
I want to innovate in [INDUSTRY/DOMAIN].
Relevant trends:
- [Trend 1]
- [Trend 2]
- [Trend 3]
For each trend:
1. How could this fundamentally change our space?
2. What new opportunities does it create?
3. What current practices will it make obsolete?
Then: What happens when we combine these trends?
Competitor Disruption
If I were a competitor trying to disrupt [MY COMPANY/PRODUCT]:
Current offering: [DESCRIBE]
Apparent weaknesses: [KNOWN WEAKNESSES]
1. Where would I attack?
2. What would I build to take customers?
3. What pricing model would undercut you?
4. What segment would I capture first?
5. What marketing message would I use?
Now: How do we do this to ourselves before someone else does?
Content Creation Prompts
Story Ideation
I need ideas for a [STORY TYPE - blog post series, video series, podcast, etc.].
Topic area: [BROAD TOPIC]
Audience: [WHO WILL CONSUME THIS]
Unique angle we have: [OUR EXPERTISE/PERSPECTIVE]
Generate:
- 10 evergreen topic ideas
- 5 timely/trending angles
- 3 controversial takes (that we could defend)
- 2 series concepts (multi-part content)
Title/Hook Generation
I'm creating content about: [TOPIC]
Core message: [WHAT I WANT TO SAY]
Audience: [WHO I'M SPEAKING TO]
Generate:
- 10 headline options
- 5 hook/opening lines
- 3 frameworks to structure the piece
Make them specific and avoid generic patterns.
Visual Concept Ideation
I need visual concepts for [PROJECT].
Purpose: [WHAT IT'S FOR]
Brand personality: [DESCRIBE BRAND]
Must convey: [KEY MESSAGES]
Generate 5 distinct visual directions:
- Concept description
- Key visual elements
- Color mood
- Photography/illustration style
- What makes it unique
Product Ideation Prompts
Feature Brainstorm
We're adding features to [PRODUCT].
Current product: [DESCRIBE]
Target users: [WHO]
Competitive gap: [WHAT WE'RE MISSING]
Generate feature ideas in categories:
1. Quick wins (low effort, high impact)
2. Differentiators (unique to us)
3. Table stakes (must-haves we lack)
4. Moonshots (ambitious bets)
For each, one line: what it is and why it matters.
New Product Concepts
Explore new product ideas in [SPACE].
Market context: [DESCRIBE THE MARKET]
Our strengths: [WHAT WE'RE GOOD AT]
Unmet needs we've observed: [CUSTOMER PROBLEMS]
Generate 10 product concepts:
- Name (can be playful/temporary)
- One-sentence description
- Target customer
- Key value proposition
- Why we could execute this
Naming Brainstorm
I need name ideas for [PRODUCT/FEATURE/COMPANY].
It should convey: [KEY ATTRIBUTES]
Tone: [SERIOUS / PLAYFUL / TECHNICAL / ETC.]
Considerations: [DOMAIN AVAILABILITY, TRADEMARK, ETC.]
Generate:
- 10 invented words
- 10 real-word names
- 5 combination/compound names
- 5 acronym options
For each, note the feeling/association it creates.
Making Ideas Actionable
Idea Evaluation
Evaluate this idea against criteria:
Idea: [DESCRIBE IDEA]
Criteria:
- Feasibility (can we do it?)
- Impact (does it matter?)
- Fit (does it align with our strategy?)
- Timing (is now right?)
- Resources (do we have what we need?)
Score 1-5 on each. Identify the biggest risk.
Rapid Prototyping Guide
I want to test this idea quickly:
[IDEA]
Design a minimum viable test:
1. What's the core hypothesis?
2. What's the smallest thing we can build to test it?
3. What would success look like?
4. What would failure look like?
5. How long should we run the test?
6. What do we do next if it works? If it doesn't?
Pitch Preparation
I need to pitch this idea:
[IDEA]
Audience: [WHO YOU'RE PITCHING TO]
Their concerns: [WHAT THEY CARE ABOUT]
Time available: [PITCH LENGTH]
Create:
1. Opening hook (15 seconds)
2. Problem statement (30 seconds)
3. Solution overview (1 minute)
4. Why now? (30 seconds)
5. Ask (what you need)
Also suggest 3 questions they might ask and how to answer.
Frequently Asked Questions
How do I get less generic AI ideas?
Include specific context, constraints, and examples. Generic prompts = generic ideas. The more specific your input, the more specific and useful the output. Our guide on how to write better AI prompts covers techniques for getting more targeted results.
Should I use AI ideas directly?
Use AI as a starting point. The best ideas often come from your reaction to AI suggestions—what you disagree with, want to modify, or get inspired by.
How many ideas should I generate?
Enough that the obvious ideas are used up and you start seeing unfamiliar ones. That number varies by problem. In practice, asking for a few dozen ideas with range constraints, then removing duplicates, is a reasonable starting point.
Can AI be truly creative?
AI remixes and recombines; humans create meaning. AI is excellent for generating raw material; humans add the judgment about what's valuable.
What if my team resists AI brainstorming?
Position it as a tool, not a replacement. Show how it accelerates the process. Let skeptics verify that human judgment remains essential.
For ready-made templates for specific creative tasks, such as product naming or story plots, browse the creative prompt templates. Run their output through the converge and select stages above.
Related Articles
AGENTS.md vs CLAUDE.md: Which AI Coding Tools Read Which Instruction File
A cross-tool guide to AGENTS.md: how Cursor, Claude Code, Gemini CLI, Kiro, Devin Desktop (Windsurf), GitHub Copilot and Antigravity read it, how it relates to CLAUDE.md, GEMINI.md and tool-specific rules, and a shared-file pattern with caveats.
How to Create Claude Code Skills: A SKILL.md Guide with a Worked Example
Create an Agent Skill for Claude Code: directory layout, SKILL.md frontmatter rules, supporting files and scripts, where Claude Code discovers skills, how they are invoked, failure handling, and how skills differ from MCP servers.
How to Evaluate AI Prompts: A Repeatable Regression Workflow
A step-by-step method for evaluating AI prompts: build a test set, write a rubric, combine deterministic and model-graded checks, record model settings, read failures and iterate without regressions.