AI for Outlining and Brainstorming
AI's strongest use case in SEO writing is not drafting — it's ideation and structure. When you ask AI to brainstorm topic angles, generate outline variations, or map out subtopics, the output is a starting point you refine. When you ask AI to write full articles, the output is a liability you have to fix. This lesson teaches you where AI accelerates your pre-writing workflow and where it reliably fails.
Part 1 — Where AI Excels in Pre-Writing
The AI Pre-Writing Sweet Spot
flowchart LR
A[AI Excels] --> B[Brainstorming\ntopic angles]
A --> C[Generating\nH2/H3 outlines]
A --> D[Clustering\nkeywords]
A --> E[Identifying\nPAA questions]
A --> F[Competitive\nH2 analysis]
G[AI Struggles] --> H[Original\nresearch]
G --> I[First-hand\nexperience]
G --> J[Brand voice\nconsistency]
G --> K[Accurate\nstatistics]
G --> L[Emotional\nnuance]
style A fill:#217346,color:#fff
style G fill:#8B0000,color:#fff
- Brainstorming Angles
- Outline Generation
- Subtopic Mapping
What AI does well: Generates 10–20 angle options in seconds. Covers angles you might miss.
What it does poorly: Cannot evaluate which angle is defensible or feasible — that requires your judgment.
The prompt pattern:
I'm writing about "[keyword]" for [audience]. The top 5 competitors all cover [common angles]. Generate 15 unique angles I could take — 5 contrarian, 5 data-driven, 5 audience-specific. For each, provide: angle title, one-sentence pitch, and difficulty to execute (easy/medium/hard).
What to do with the output: Immediately discard 80%. Keep the 2–3 angles that are both differentiated and feasible for you to execute with your available data/experience.
What AI does well: Creates structurally sound H2/H3 outlines that cover the topic comprehensively.
What it does poorly: Defaults to generic structures ("Introduction," "Benefits," "Conclusion"). Over-covers breadth at the expense of depth.
The prompt pattern:
Create an article outline for "[keyword]" with:
- H1: Must contain keyword + outcome promise
- 6–8 H2s: Each must pass the skim test (descriptive, not generic)
- 2–3 H3s per H2 as needed
- NO "Introduction" or "Conclusion" headings
- Target audience: [audience]
- Differentiation angle: [your angle]
What to do with the output: Merge the best parts of 2–3 generated outlines. Cut sections that add breadth but not depth. Add sections from your competitor analysis gaps.
What AI does well: Exhaustively maps related subtopics and questions that readers might have.
What it does poorly: Cannot prioritize which subtopics are worth covering vs. which are tangential.
The prompt pattern:
For the topic "[keyword]," map every question a [audience] would ask. Organize into: Must-Know (include in main article), Nice-to-Know (include as H3 or FAQ), and Out-of-Scope (separate article). Base this on common knowledge gaps, not just keyword tools.
Part 2 — The Outlining Workflow
From Brief to Approved Outline (with AI)
flowchart TD
A[Receive SEO Brief] --> B[Apply 3-Question Filter\nHuman Only]
B --> C[AI: Generate 3\nOutline Variations]
C --> D[Human: Merge Best\nParts of Each]
D --> E[AI: Expand Each H2\nwith Suggested Talking Points]
E --> F[Human: Cut Irrelevant\nPoints + Add Expert Insights]
F --> G[Human: Verify Against\nCompetitor H2 Matrix]
G --> H[Submit Outline\nfor Approval]
style B fill:#F4A261,color:#000
style D fill:#F4A261,color:#000
style F fill:#F4A261,color:#000
style G fill:#F4A261,color:#000
style C fill:#264653,color:#fff
style E fill:#264653,color:#fff
AI generates 60% of the raw material. You refine, filter, and add 40% — including angle decisions, competitor-gap sections, and brand-specific knowledge. The human 40% is where all the differentiation lives.
Part 3 — Prompt Architecture for Outlining
The Structure of Effective Prompts
- Prompt Anatomy
- Full Example Prompt
Every effective outlining prompt has 5 components:
| Component | Purpose | Example |
|---|---|---|
| Role | Sets the AI's expertise level | "You are a senior SEO content strategist." |
| Task | What you want it to produce | "Create an article outline for [keyword]." |
| Context | Background the AI needs | "Target audience: [who]. Angle: [what]. Competitors cover: [topics]." |
| Constraints | Rules that prevent common failures | "No 'Introduction' heading. Max 8 H2s. Each H2 must be descriptive." |
| Output format | How to structure the response | "Format: H2 → 2–3 H3s → 3 bullet-point talking points per H3." |
Role: Senior SEO content strategist with 10 years of experience in B2B SaaS.
Task: Create 3 outline variations for an article targeting the keyword "email marketing automation for small business."
Context:
- Target audience: Small business owners (1–20 employees), non-technical
- Differentiation angle: "We tested 5 automation tools for 60 days with a real team"
- Competitors cover: tool lists, feature comparisons, generic "getting started" guides
- Missing from competitors: real test results, cost breakdowns, setup time data
Constraints:
- No "Introduction" or "Conclusion" headings
- Each H2 must pass the skim test — a reader reading only H2s should understand the full article
- 6–8 H2s per outline
- Include an FAQ section targeting PAA questions
Output format: For each variation, provide:
- H1 title
- H2 headings with 2–3 H3s each
- 3 talking points per H3 (bulleted)
Part 4 — Bad vs. Good Examples
- ❌ Lazy Prompting
- ✅ Strategic Prompting
Prompt: "Write an outline for an article about email marketing."
AI Output:
1. Introduction
2. What Is Email Marketing?
3. Benefits of Email Marketing
4. How to Get Started
5. Best Practices
6. Tools to Use
7. Conclusion
(Why it fails: The prompt gave AI no context, no audience, no angle, no constraints. The output is a generic template that matches 500 other articles. It provides zero competitive advantage.)
Prompt: [Uses the full 5-component structure above]
AI Output (1 of 3 variations):
H1: We Tested 5 Email Automation Tools for 60 Days — Here's What Actually Works for Small Teams
H2: Why Most "Best Of" Lists Are Useless for Small Businesses
H3: The Enterprise Bias in Tool Recommendations
H3: What "Affordable" Actually Means When You Have 500 Subscribers
H2: Our Testing Methodology (So You Know We're Not Making This Up)
H3: The 5 Tools We Tested and Why
H3: What We Measured: Setup Time, Cost, Deliverability, Support
H2: The Results: Ranked by What Matters Most
H3: Best Overall for Small Teams
H3: Best on a Budget
H3: Best for Automation Power Users
H2: The Setup Nobody Tells You About
H3: DNS Configuration and Authentication
H3: List Migration Without Losing Subscribers
H2: FAQ: What We Get Asked After Publishing This
(Why it wins: The outline has a clear angle (testing), specific audience (small teams), and differentiated structure (results-based, not feature-based). Every H2 passes the skim test.)
Part 5 — AI Pitfalls to Watch
The Breadth Trap
AI produces outlines that try to cover everything. 12 H2s, each with 4 H3s. The result is 4,000 words of surface-level coverage. Fix: Limit to 6–8 H2s. Cut any H2 that doesn't serve your differentiation angle.
The "Introduction" Reflex
AI always includes "Introduction" and "Conclusion" as headings — even when told not to. Fix: Delete them every time. Your H1 IS your introduction. Your final section IS your conclusion.
The Duplicate Angle Problem
When generating multiple outline variations, AI produces 3 versions that are 80% identical. Fix: Force differentiation: "Variation 1: beginner-focused. Variation 2: data-driven. Variation 3: contrarian."
Part 6 — Output Checklist
- Pre-writing focus: You use AI for brainstorming and outlining, not for skipping the thinking step.
- 5-component prompts: Your prompts include Role, Task, Context, Constraints, and Output Format.
- Multiple variations: You generate 2–3 outline options and merge the best parts.
- 60/40 ownership: AI generates raw material; you add angle, gaps, and expert knowledge.
- No generic headings: Your final outline has zero "Introduction," "Conclusion," "Overview," or "Tips" headings.
- Competitor-validated: Your final outline covers competitor H2 gaps identified in your SERP analysis.
Internal use only. Do not distribute externally. For questions or suggested updates, raise with the content lead.