The trick is to think of ChatGPT as a brainstorming partner instead of an idea vending machine. You think, and let ChatGPT handle the volume and structure of ideas. Over time, I am experimenting with a small set of prompts that expand clarity and range for my ideas.
I can’t claim to have developed them, as there are umpteen examples all around, especially on Reddit. It’s up to us to test this compilation on our individual projects. The goal is to turn fuzzy thoughts into testable ideas before we get down to the practical stuff. For this article, I am trying a brainstorming exercise on how I could design my newsletter to stay interesting long-term.
1
Help ChatGPT understand your context
Set expectations early to avoid generic ideas
Before generating ideas, I always start by telling ChatGPT how to respond. This “core” prompt frames the entire session and instructs the model to avoid surface-level suggestions.
You are my structured brainstorming partner. Goal: Help me brainstorm around [describe topic/problem/output]. Context: [who it’s for, constraints, timeline, resources]. Process: Ask up to 5 clarifying questions one by one. Propose 3–5 different brainstorming frameworks we could use (e.g., SCAMPER, mind map, assumptions, constraints, first principles). After I choose one, guide the brainstorming in short, numbered lists, aiming for specificity and practicality. Ultimately, cluster the best ideas into 3–5 themes and suggest concrete next steps. Write in concise bullet points and avoid generic advice.
The goal and context go upfront, so the model understands your situation. I’ve found this prompt especially useful for complex work like content planning or life goal strategy. Starting with this detailed back and forth here saves a lot of effort later. Remember, what matters is context, clarity, and intent for all prompts.
2
Generate volume first with classic divergent ideas
Speed and range matter more than quality at the start
This is a classic from Brainstorming 101. When you’re stuck, the fastest way forward is often to generate as many ideas as you can. Even stupid ones. This prompt forces ChatGPT and your brain to grease the gears by asking for volume first. Culling them down and prioritization can come later.
Let’s do divergent brainstorming on [topic]. Constraints: [budget/time/industry/audience/etc.]. Step 1: Generate 20 quick, rough ideas in a numbered list, from obvious to wild. Step 2: For each idea, add one short phrase on why it might work. Step 3: Mark the five ideas that seem most promising for impact vs effort.
The last step is critical. Without it, you end up with a list you never revisit. A numbered list also makes it easy to reference specific ideas in follow-up prompts. The best ideas usually can be found in the latter part of the list. Try combining ChatGPT with a mindmap to find associations between ideas.
3
Use reverse brainstorming to expose hidden risks
Break ideas to see what actually matters
Reverse brainstorming flips the problem: instead of asking how to succeed, you ask how to fail. It’s like carrying out a pre-mortem before the event.
Use reverse brainstorming for [topic]. List 15 ways to make this problem worse. For each “worse” idea, flip it into a constructive idea or safeguard. Present the flipped list as concrete, testable actions or concepts.
This prompt surfaces weak ideas and common mistakes fast. If an idea fails because it’s too generic, going niche becomes your lever. If it fails due to complexity, simplicity is the answer. I use this when an idea feels “fine” but not compelling and it usually explains why.
4
Take leaps of imagination with random trigger words
Unrelated inputs often unlock non-obvious directions
In an analog world, I carried out this exercise with a dictionary or a newspaper. When everything starts to sound the same, randomness helps. This prompt forces unrelated inputs to come together with new connections. But be warned, this is really tough on the brain cells… perhaps that’s why ChatGPT is such a help.
I want creative, non-obvious ideas for [topic]. Step 1: Generate five random, unrelated trigger words. Step 2: For each word, show three ways it could inspire or modify my idea. Step 3: Summarize the five most interesting directions that emerged.
The trigger words themselves don’t matter. The value comes from translating them into metaphors, constraints, or perspectives you wouldn’t normally consider.
5
Think out of the box with the SCAMPER framework
Frameworks help you follow a thinking plan
Edward de Bono popularized lateral thinking. SCAMPER is a classic creativity framework for doing it. and it’s blazingly fast with AI.
Use the SCAMPER framework to brainstorm around [product/project/topic]. For each letter: Substitute, Combine, Adapt, Modify, Put to other uses, Eliminate, Reverse. Generate 3–5 specific ideas, not generic advice. Keep each idea to one sentence. End with a short list of the top 5 ideas across all SCAMPER steps.
This prompt works really well with a good base idea. So, I would use it much later in the brainstorming process. It helps you see how small changes like removing a feature or combining two elements can dramatically improve clarity or usefulness.
6
Pressure-test ideas with a SWOT-style analysis
Strategic thinking improves idea quality and execution
A SWOT-style prompt pushes brainstorming into the real world and highlights the trade-offs you might have to make down the road.
Run a SWOT-style brainstorming session on [idea or project]. List 5–7 Strengths, 5–7 Weaknesses, 5–7 Opportunities, 5–7 Threats
3 ideas that double down on strengths
3 ideas that fix or bypass weaknesses
3 ideas that capture opportunities
3 ideas that defend against threats
Present everything in concise bullets.
I often use this prompt when I don’t want to gloss over an idea that sounds great on paper (or screen). Using a visual diagramming tool for brainstorming can help you get a bird’s-eye view.
7
Shift perspectives with rolestorming
Different viewpoints reveal needs you didn’t consider
Role storming forces you to be in someone else’s shoes. This prompt is invaluable for finding pain points with a product or a process.
We’ll do “rolestorming” for [topic]. Take the perspective of: – A beginner – A power user – A critical stakeholder – A competitor
For each persona, list five ideas, requests, or improvements they would push for. Then merge the best ideas into one prioritized list.
As a writer, overlapping insights across personas help me finetune my content for a wider readership.
8
Convert ideas into actionable roadmaps
Turn brainstorms into practical plans
Close your brainstorming with this prompt:
Take the list of ideas I provide and turn them into a practical action roadmap.
Score each idea on impact (1–5) and effort (1–5) with a brief explanation.
Place them into Quick wins, Long bets, Experiments, and Cancel.
Suggest a 2-week action plan with specific tasks to test 3 of the best ideas.
You don’t want your brainstorms to end up in a document that doesn’t go anywhere. The best ideas shouldn’t be based on gut feeling alone. I often feel the best ones are a balance between instinct and some hard proof.
Combine two prompts and see how your ideas change
Productive brainstorming isn’t about having better ideas. It’s about understanding your own blind spots, too. You don’t need to use all the prompts (though that too can help). Start with the core setup prompt, then pair one divergent prompt with a litmus test prompt like the SWOT analysis. The jump in clarity can be revealing.
