I used to dread content planning days. You know the feeling—staring at a blank document, waiting for inspiration to strike while your cursor blinks mockingly. Some weeks I'd spend entire afternoons generating maybe five mediocre ideas. Other weeks? Complete creative block.
Then I discovered that I'd been asking AI the wrong questions entirely.
The Mindset Shift That Changed Everything
Here's what most people get wrong about AI brainstorming: they treat it like a vending machine. Pop in "give me blog ideas about productivity" and expect golden nuggets to fall out. Spoiler alert—you'll get the same generic suggestions everyone else gets.
The breakthrough came when I stopped asking AI to generate ideas and started using it as a thinking partner.
Think of it like this: AI has read millions of articles, analyzed countless patterns, and absorbed more information than any human could. But it doesn't know your audience, your unique angle, or what's already working for you. When you combine your context with its pattern-matching abilities? That's where the magic happens.
My Actual Workflow (Step by Step)
Let me walk you through exactly what I do every Monday morning. No theory—just the messy, practical reality.
Phase 1: Setting the Stage
Before I type a single prompt, I gather three things:
- My top 5 performing posts from the past month (titles + why I think they worked)
- Three pain points I've heard from readers recently
- One trend I've noticed in my space
Then I open Claude and start with what I call my "context dump":
I'm a content creator focused on making AI accessible to beginners and
business professionals. My audience ranges from curious newcomers to
small business owners exploring AI tools.
Here are my 5 best-performing articles recently:
- "What Is Generative AI?" (explained complex concept simply)
- "Getting Started with ChatGPT" (actionable, beginner-friendly)
- "Prompt Engineering Basics" (practical tips people could use immediately)
- "What Are Embeddings?" (made technical concept relatable)
- "Building Your First Zapier Agent" (step-by-step walkthrough)
Pain points I keep hearing:
- "I don't know where to start with AI"
- "Everything I read is too technical"
- "I want practical examples, not theory"
Current trend: More people asking about AI agents and automation.
This context transforms every suggestion that follows. Instead of generic "10 AI Tips" ideas, I get suggestions that actually fit my voice and audience.
Phase 2: The First Brainstorm Burst
Now comes the fun part. I ask for quantity over quality:
Based on this context, generate 25 blog post ideas that:
- Address at least one of those pain points
- Could work as beginner-friendly guides OR practical tutorials
- Offer a fresh angle I haven't covered
For each idea, give me:
- A working headline
- One sentence on why this matters to my audience
- What makes this different from typical coverage of the topic
Within about 30 seconds, I have 25 ideas to work with. Are they all winners? Absolutely not. But that's the point—I'm mining for gems, not expecting perfection.
Phase 3: The Iteration That Makes It Work
Here's where most people stop. They take those 25 ideas and call it a day. But the real value comes from iteration.
I scan the list and pick my top 5-7 favorites. Then I go deeper:
I like these ideas best:
- #3: "Why Your First AI Project Should Be Boring"
- #7: "The 5-Minute AI Test That Reveals If a Tool Is Worth Your Time"
- #12: "What Nobody Tells You About Prompt Engineering"
For each of these, brainstorm 5 different angles I could take:
- A contrarian angle
- A personal story angle
- A "common mistakes" angle
- A step-by-step tutorial angle
- A "lessons learned" angle
Suddenly, my 3 favorite ideas become 15 fully-formed concepts. And because I've given Claude specific frameworks to work with, the suggestions have real depth.
The Prompts That Actually Work
After months of experimentation, I've landed on a few prompt patterns that consistently deliver:
The "Contrarian Spark" Prompt
What's a commonly accepted belief in [your niche] that might actually
be wrong or oversimplified? Generate 10 blog ideas that challenge
conventional wisdom while still being helpful and accurate.
This one produces ideas like "Why You Shouldn't Start with ChatGPT (And What to Use Instead)" or "The Productivity Myth: When AI Actually Slows You Down."
The "Gap Finder" Prompt
Look at these topics I've already covered: [list them]
What related questions might my readers have that I haven't addressed?
What natural "next steps" would someone want after reading these?
This helps me build content clusters that actually make sense together.
The "Persona Collision" Prompt
Imagine three people are brainstorming content for my site:
- A complete beginner who just heard about AI yesterday
- A small business owner who's skeptical but curious
- A tech enthusiast who wants to go deeper
What questions would each person want answered? Where do their
interests overlap?
The intersection points often reveal the most universally appealing topics.
What I've Learned the Hard Way
A few pitfalls I walked straight into (so you don't have to):
The "First Response" Trap
AI's initial suggestions are usually its most obvious ones—the ideas that show up first in search results, the angles everyone's already covered. The gems are in the follow-up questions. Push past the first response.
The Generic Spiral
If you're getting bland suggestions, your prompt is probably too vague. "Give me content ideas" will never work as well as "Give me content ideas for small business owners who are intimidated by AI but need to automate their customer support."
The Quantity Illusion
Having 100 ideas feels productive. But 100 mediocre ideas are worth less than 5 great ones that you'll actually write. I now spend more time on the iteration phase than the initial brainstorm.
Forgetting Your Voice
AI can suggest what to write about, but it shouldn't dictate how you think about topics. I always ask myself: "What's my unique take on this? What experience or insight can I add that no AI summary could provide?"
The Numbers (Because I Know You're Curious)
Since adopting this workflow:
- Brainstorming time: Dropped from 4-6 hours weekly to about 45 minutes
- Ideas generated per session: Went from 5-10 to 50-100
- Usable ideas: About 20% make it to my content calendar (but that's still 10-20 solid concepts per session)
- Creative blocks: Basically eliminated. When I'm stuck, I just iterate differently.
The time savings alone would make this worthwhile. But honestly? The bigger win is that brainstorming became fun again. It's collaborative now, not lonely.
Tools I Actually Use
I've tried most of the AI brainstorming tools out there. Here's my honest assessment:
Claude (my primary choice): Best for nuanced, contextual brainstorming. It handles long context well and pushes back thoughtfully when my prompts are vague.
ChatGPT: Great for rapid-fire generation when I need raw volume. The memory feature helps it learn my preferences over time.
Notion AI: Useful when I want to brainstorm directly in my content calendar. The integration is seamless if you're already a Notion user.
For most sessions, I stick with Claude. The quality of suggestions just feels more aligned with how I actually think.
Try This Monday
Here's a challenge: Block 30 minutes next Monday morning. Gather your context (top posts, pain points, one trend), and try the workflow I described.
Start with that context dump. Ask for 25 ideas. Pick your favorites and iterate with different angles. See what emerges.
I'd bet you'll generate more usable ideas in that 30 minutes than you have in the past month of staring at blank documents.
The goal isn't to outsource your creativity to AI. It's to give your creativity a running start.
What's Next
Now that you've got a backlog of ideas, the challenge becomes execution. In a future post, I'll share how I use AI to outline posts quickly without losing my authentic voice.
For now, go fill that content calendar. Your future self (the one who isn't scrambling for ideas at deadline time) will thank you.
Have questions about this workflow? Want to share what's working for you? Drop a comment below—I read every one.