The 3D Idea Matrix
Every great content idea has three parts: a topic, an angle, and a format. Most people only ever change the topic, which is why their ideas feel flat. This is the system I use to never run out of genuinely good ideas — not the half-baked ones AI hands you by default. Give the prompts below your niche and the creators you follow, and get 10 ideas at a time, each with the angle and format already picked.
What this is
Picture a 3D graph with three axes. The first axis is your topic — the subjects you cover in your niche. That's the easy part. The hard part is the second axis: the angle, the slant that makes a familiar topic feel novel and worth watching. The third axis is the format — how the idea actually gets shot.
Any point where those three axes meet is a content idea. Multiply your topics by 50 angles by 12 formats and you have more ideas than you could ever film. The trick isn't more brainstorming — it's a system that generates the permutations for you, then lets you pick the ones you'd actually stand behind.
A quick example: "how to lead an executive meeting" is a boring topic. But run it through the cross-domain transfer angle — borrow a concept from a completely different niche, like how you'd teach a kindergartner — and suddenly it's a video that reached over a million and a half people. Same topic. Different angle.
How to use it
Run the four prompts below one stage at a time, in the same chat, and answer each stage's questions before moving to the next. Each stage builds context for the next: your subtopics, the creators you track, cross-domain crossovers, and finally the ideas themselves. The more specific you are about the people you follow, the better the references it pulls.
Works in: Claude, ChatGPT, or any capable AI. Rule of thumb: one prompt at a time, answer in between, keep it all in a single conversation.
The prompts
Paste these one at a time, in order, in the same chat.
We're building a content map for my niche. My niche is: [WRITE YOUR NICHE IN ONE LINE] Brainstorm an exhaustive list of subtopics inside my niche — the specific things I could teach. Group them into 5–8 themes, and push for 30–50 subtopics total. Don't stop at the obvious ones; include the narrow, tactical, and slightly weird corners too. When you're done, ask me which themes to keep, cut, or add before we move on.
Now let's map the field. Here are the creators and competitors I track in my niche: [LIST 5–15 CREATORS / COMPETITORS] And here are the people I personally follow and learn from (in and out of my niche): [LIST THE KEY PEOPLE YOU FOLLOW] For each one, tell me: the topics they win with, the angles they lean on, and the formats they use most. Then show me the white space — the topics and angles NObody in that list is covering well — and which of their content structures I could borrow.
Now do cross-domain transfer. Pull concepts, frameworks, and metaphors from OUTSIDE my niche — other industries, disciplines, sports, hobbies, history — and map them onto my subtopics to create genuinely novel ideas. Give me 10 cross-domain idea seeds. For each: name the outside concept, the subtopic it maps onto, and the one-sentence idea. Favor transfers that reframe something boring in my niche as something surprising.
Now generate content ideas — 10 at a time. Use my subtopics (Stage 1), the white space and borrowed structures (Stage 2), and the cross-domain seeds (Stage 3), applied through the 50 angles below. For each idea give me: - Topic - Angle (name it) - The idea in one specific sentence - A recommended format: Yap, Green Screen, Talking Head, Split Screen, Whiteboard, Interview, Object Lesson, Play/Pause Reaction, Clone, Ask Me Anything, Listicle, Carousel Keep them specific and non-obvious. After each set of 10, ask if I want the next 10 or want to expand any idea into a full outline. THE 50 ANGLES: 1 Actionable step-by-step · 2 Copy-paste system · 3 Speedrun · 4 Pre-flight checklist · 5 Teardown of your own process · 6 Tool walkthrough/demo · 7 Automation build · 8 Fix-this-mistake · 9 Contrarian take · 10 Stop doing X · 11 Overrated/underrated · 12 Myth-bust · 13 Unpopular opinion you'll defend · 14 The obvious answer is wrong · 15 Permission-to-not · 16 X vs Y · 17 Before/After · 18 Good vs bad example · 19 Persona A vs B · 20 Old way vs new way · 21 What I thought vs what's true · 22 Cheap vs expensive · 23 Cross-niche concept transfer · 24 Metaphor transfer · 25 Steal-from-industry · 26 First-principles from another discipline · 27 Historical analogy · 28 "This is just like ___" · 29 Origin story of a build · 30 Failure/lesson · 31 Highs & Lows · 32 Real-time build log · 33 Behind-the-scenes · 34 Milestone/receipt · 35 Day-in-the-life of the workflow · 36 Teardown of someone else's work · 37 Pattern-spotting · 38 Framework introduction · 39 Mental model · 40 Decision tree/diagnostic · 41 Glossary/decode · 42 The 80/20 · 43 Q&A/AMA · 44 Objection handling · 45 Common-pain callout · 46 Prediction/future · 47 Observation · 48 Listicle · 49 Resource drop · 50 Challenge/experiment
Get a new build-with-AI guide every week
Weekly playbooks on prompting, agents, and content systems — free.
You're in. First email lands shortly.
The 50 angles
The second axis. These are the slants I cycle through to make any topic feel new. They're already baked into the prompts above — this is the full list for reference. The Cross-Domain Transfer and Story families tend to produce the most original ideas.
Tactical / How-To
- Actionable step-by-step
- Copy-paste system
- Speedrun
- Pre-flight checklist
- Teardown of your own process
- Tool walkthrough / demo
- Automation build
- Fix-this-mistake
Contrarian / Reframe
- Contrarian take
- Stop doing X
- Overrated / underrated
- Myth-bust
- Unpopular opinion you'll defend
- The obvious answer is wrong
- Permission-to-not
Comparison / Contrast
- X vs Y
- Before / After
- Good vs bad example
- Persona A vs B
- Old way vs new way
- What I thought vs what's true
- Cheap vs expensive
Cross-Domain Transfer
- Cross-niche concept transfer
- Metaphor transfer
- Steal-from-industry
- First-principles from another discipline
- Historical analogy
- "This is just like ___"
Story / Experience
- Origin story of a build
- Failure / lesson
- Highs & Lows
- Real-time build log
- Behind-the-scenes
- Milestone / receipt
- Day-in-the-life of the workflow
Analysis / Teach
- Teardown of someone else's work
- Pattern-spotting
- Framework introduction
- Mental model
- Decision tree / diagnostic
- Glossary / decode
- The 80/20
Audience / Curiosity
- Q&A / AMA
- Objection handling
- Common-pain callout
- Prediction / future
- Observation
- Listicle
- Resource drop
- Challenge / experiment
The 12 formats
The third axis — how the idea actually gets shot. These are the twelve I keep coming back to. Match the format to the idea: demos want Split Screen, frameworks want Whiteboard, hot takes want Yap.
Talk straight to camera, casual, no visuals. Live or die by the hook and one sharp idea.
You in front of an article, post, or your own notes — reveal points one at a time.
Studio-style direct address with light b-roll. Your reliable default.
Your face plus an on-screen demo. Best for tool walkthroughs and tech tutorials.
Draw a concept or the steps as you explain. Great for frameworks and processes.
Read a supplied question and give your raw, unscripted take.
Use physical objects to stand in for an idea and rearrange them live.
Play a viral clip and pause to break down, stage by stage, why it works.
Two of you on screen — before/after, good vs bad, or two personas in conversation.
Answer a real audience question. Conceptual, no show-and-tell needed.
A ranked or numbered list — or a reveal of options one at a time — carried by an on-screen tally.
A swipeable multi-slide post (not video) for step-by-step breakdowns and frameworks.