10 AI Prompts for Blog Outlines: A Research-to-Outline Library

T
Tommy
11 min readNov 6, 2026Updated Nov 6, 2026
A tailor's hands marking a chalk line along a wooden ruler on blue wool cloth with pattern outlines drawn out and steel shears lying beside it

The outline phase is where a lot of posts stall. You know the topic, but turning research into a structured, search-aware outline is a separate skill, and generic prompts make it worse by producing cookie-cutter structures that ignore what your audience is actually asking.

This article gives you a library of AI prompts for blog outlines, arranged in research-to-outline order. Each one comes with a filled-in example input, an output snippet, and a "fix by hand" note covering what the AI tends to get wrong. All ten use the same sample topic, email deliverability for small SaaS, so you can see how one output feeds the next. They work in ChatGPT, Claude, Gemini or any capable model. If your outputs also need to sound like you, pair this library with a brand voice guide for AI writing.

Quick Answer: The most effective AI prompts for blog outlines include your target audience, the search intent you're addressing, the questions to answer, and the tone you want. Use them as a sequence, from audience research through SERP gap analysis to title variants, and edit every output by hand.

How to Use This Prompt Library

Run the prompts in order inside one chat session so the AI keeps context: the gaps from Prompt 3 feed the outline in Prompt 5 without you repeating them. Replace every square-bracket placeholder such as [topic] before running. All example outputs are illustrative, not real data.

Treat every output as a draft. Names, numbers and claims need a fact-checking pass before they reach a post.

10 AI Prompts for Blog Outlines, From Research to Ready

1. Audience Research

Prompt: You are a content strategist. For a blog post about [topic], describe 2-3 distinct reader personas. For each, include: job role, biggest pain point related to [topic], what they already know, and what they still need to learn.

Example input: Topic = email deliverability for small SaaS; audience = early-stage SaaS founders and their marketing leads.

Example output snippet (illustrative example only):

Persona 1 — SaaS Founder (non-technical): Understands the product deeply,
struggles with why emails land in spam. Knows Mailchimp exists. Doesn't
know what SPF/DKIM records do.
Persona 2 — Junior Marketing Lead: Sends campaigns regularly, tracks
open rates, but hasn't audited sender reputation.

Naming each persona's specific knowledge gap stops you writing for a vague "beginner."

Fix by hand:

  • Replace invented personas with what you know from real customers, support tickets or sales calls.
  • Cut any persona you could not name a real person for.

2. Audience Objections and Questions

Prompt: Based on the audience for a blog post about [topic], list the 8 most common objections, doubts, or questions they have before they take action. Format as direct questions they would ask.

Example input: Audience = early-stage SaaS founders worried about email deliverability.

Example output snippet (illustrative example only):

1. "Is this something I need to fix now, or only at scale?"
2. "Can I improve deliverability without touching DNS settings?"
3. "How do I know if my emails are landing in spam?"
4. "Will switching ESPs help, or is the problem on my end?"

This list becomes your FAQ section and a source for H2 headings.

Fix by hand:

  • Merge near-duplicates, then check the list against real questions from Reddit threads, support inboxes or "People also ask."
  • Rewrite stiff questions into the way your readers actually phrase them.

3. SERP Gap Mining

Prompt: Act as an SEO analyst. For the keyword [primary keyword], list 6-8 subtopics or angles that commonly-ranking articles tend to ignore or underserve. Focus on [audience type]'s specific concerns. Label each gap as: Missing, Shallow, or Outdated.

Example input: Keyword = email deliverability for SaaS; audience = non-technical founders.

Example output snippet (illustrative example only):

Gap 1 (Missing): How deliverability differs for transactional vs.
marketing email in early-stage SaaS
Gap 2 (Shallow): What a healthy sender reputation looks like for a
small list
Gap 3 (Outdated): Open-rate tracking after Apple's Mail Privacy
Protection changes

This is the highest-leverage prompt in the library, but the AI is guessing from memory, not reading today's results page.

Fix by hand:

  • Open the live top 5 results and confirm each "gap" is really missing. Delete the ones that are not.
  • Our content gap analysis guide covers free and paid ways to check, and free keyword research helps you choose the keyword to feed in.

4. Search Intent Check

Prompt: For the keyword [keyword], classify the dominant search intent (informational, navigational, commercial, or transactional). Then describe what a searcher at this stage wants to accomplish, what format would satisfy them best, and what they do NOT want to see in a result.

Example input: Keyword = "how to improve email deliverability SaaS."

Example output snippet (illustrative example only):

Intent: Informational (with commercial undertones)
Searcher goal: Diagnose and fix a deliverability problem without
hiring a specialist
Best format: Step-by-step guide with diagnostic checkpoints
What they don't want: Generic ESP comparisons, enterprise advice

If the intent is diagnostic, a "10 best tools" listicle will underperform however well it is written. Our guide to search intent SEO covers the full method.

Fix by hand:

  • Search the keyword yourself and compare the AI's verdict to the formats that actually rank.
  • If the results disagree with the AI, trust the results page.

5. Outline from Gaps

Prompt: Using the content gaps [paste gaps from Prompt 3] and the search intent [paste from Prompt 4], create a detailed blog post outline for [topic]. Include an H1, 6-8 H2 sections, and 2-3 H3 sub-points per section. Prioritize the gaps as early H2s where possible.

Example input: Topic = email deliverability for small SaaS; gaps and intent pasted from the previous prompts.

Example output snippet (illustrative example only):

H1: Email Deliverability for Small SaaS: A Non-Technical Fix-It Guide
H2: Why Deliverability Works Differently at Early Stage
  H3: Transactional vs. marketing email: why the rules split
H2: Diagnosing Your Sender Reputation (Before You Change Anything)
  H3: Free tools to check your standing today

Because personas, gaps and intent come first, the outline reflects real reader needs rather than the AI's default idea of a deliverability post.

Fix by hand:

  • Cut or merge any H2 that repeats another, and reorder so the reader's first question is answered first.
  • Add the one section only you can write: your own data, a client story or a screenshot-worthy example.

6. Section-Level Expansion

Prompt: Expand the section "[H2 section title]" from this outline: [paste outline]. Add 4-5 specific talking points, one analogy or example, one potential stat or data hook (note if unverified), and a suggested word count for this section.

Example input: Section = "Diagnosing Your Sender Reputation"; outline pasted from Prompt 5.

Example output snippet (illustrative example only):

Talking points:
- How to read a Google Postmaster Tools dashboard
- Why a new sending domain needs a gradual warm-up
- Checking SPF, DKIM and DMARC before blaming content
Analogy: Sender reputation works like a credit score
Suggested word count: 350-400 words

Expansion stops you writing hollow H2s that exist only to satisfy a structure.

Fix by hand:

  • Treat every stat hook as unverified until you find the primary source. If you cannot, drop it.
  • Replace the generic analogy with one from your own experience if you have one.

7. Writing the Introduction

Prompt: Write three alternative introductions for a blog post titled "[title]" targeting [audience]. Each intro should use a different hook style: (1) a surprising statistic or counterintuitive fact, (2) a relatable frustration scenario, (3) a direct challenge to conventional wisdom. Keep each under 80 words.

Example input: Title = "Email Deliverability for Small SaaS: A Non-Technical Fix-It Guide"; audience = non-technical SaaS founders.

Example output snippet (illustrative example only):

Hook 1 (Counterintuitive): Your emails can have a perfect design and
a strong subject line and still never reach the inbox.
Hook 2 (Scenario): You hit send on your onboarding sequence. Three
days later, a trial user says the welcome email never arrived...

Three options let you make an editorial choice instead of accepting the first draft. The frustration scenario suits readers already living the problem.

Fix by hand:

  • The "statistic" hook is the one models most often invent. Keep it only with a real, sourced number.
  • Cut stock openers like "In today's digital landscape" and rewrite the first sentence in your own voice.

8. Objection Handling

Prompt: For a blog post about [topic] targeting [audience], take these 3 audience objections: [paste from Prompt 2]. For each, write a 2-3 sentence response that acknowledges the concern, provides a concrete counter-point, and suggests a first action step.

Example input: Objections from Prompt 2; topic = email deliverability for small SaaS.

Example output snippet (illustrative example only):

Objection: "Is this something I need to fix right now?"
Response: Reputation problems are easier to reverse before they
compound, and high spam complaint rates can trigger filtering.
First step: Check Google Postmaster Tools before your next send.

Answering a doubt at the moment it appears keeps a reader from bouncing.

Fix by hand:

  • Verify every concrete claim, such as thresholds or "this will trigger filtering," against the provider's own documentation.
  • Make the first action step something a reader can do in ten minutes, not "audit your strategy."

Prompt: Given a blog post about [topic] on a site that covers [your site's content areas], and this list of existing pages: [paste titles and URLs], suggest 5-7 internal linking opportunities. For each, provide the anchor text, the target page, and the section in the current post where the link fits naturally.

Example input: Topic = email deliverability for SaaS; site covers email marketing, SaaS growth and product onboarding; a pasted list of your published posts.

Example output snippet (illustrative example only):

Link 1: Anchor "onboarding email sequence" → your onboarding
email guide → place in: transactional emails section
Link 2: Anchor "set up DKIM and SPF" → your authentication
tutorial → place in: diagnosis section

Planning links at outline stage means every section is built with your site structure in mind. Our guide to topic clusters and internal linking explains how to choose targets.

Fix by hand:

  • Open every suggested URL. Models invent plausible pages, so only link to ones that exist and load.
  • Rewrite anchors that are stuffed with the exact keyword so they read naturally in the sentence.

10. Title and Meta Variants

Prompt: Generate 5 title options and 3 meta description options for a blog post about [topic] targeting [audience]. Titles should vary by format (question, number, how-to, bold claim). Meta descriptions should be under 160 characters and include a clear benefit statement.

Example input: Topic = email deliverability for small SaaS; audience = non-technical founders.

Example output snippet (illustrative example only):

Titles:
1. "Email Deliverability for SaaS: Fix It Without Touching DNS"
2. "7 Reasons Your SaaS Emails Land in Spam (And How to Fix Each)"
3. "Is Your Email Deliverability Hurting Your Trial Conversions?"
Meta: "Struggling with emails landing in spam? This non-technical
guide shows SaaS founders how to diagnose and fix deliverability."

Fix by hand:

  • Count characters yourself. Models regularly misreport length, and a title over about 60 characters risks truncation in results.
  • Never promise something the post does not deliver, such as "without touching DNS" if one section requires it. Our guide to writing a meta title and meta description that get clicks has the checklist.

Getting More From the Library

Try role-stacking: give the AI both an expertise role and a communication-style role, for example "an SEO analyst who explains things to non-technical founders the way a patient CTO would." The style constraint cuts jargon-heavy output.

Chain a critique loop after Prompt 5: ask the AI to name the weakest section of the outline and how to strengthen it. It takes seconds and often catches structural gaps before you draft a word.

Frequently Asked Questions About AI Prompts for Blog Outlines

What makes an AI prompt effective for creating blog outlines? It includes your target audience, the search intent of your keyword, the content gaps you want to address, and the structure you want back. More relevant context gives more usable output.

Can I use these prompts in ChatGPT, Claude, or other AI tools? Yes. They are model-agnostic. Run them in sequence within one session so the AI retains context from earlier prompts.

Should I publish AI-generated outlines without editing them? No. Review for brand voice, verify factual claims, add your own data and experience, and check that the structure matches what readers need.

How many prompts do I need for a complete outline? A minimal sequence uses three: audience research, search intent check and outline generation. For competitive keywords, run all 10.

Conclusion

A structured prompt sequence turns AI from a vague writing assistant into a research and planning tool, as long as you edit what it gives you. The "fix by hand" step is where the outline stops being generic and starts being yours.

Key Takeaways:

  • Run prompts in sequence within one chat so the AI retains context
  • Always include audience specifics and content gaps in your prompts
  • Verify SERP gaps, stats, URLs and character counts yourself; these are the most common failure points
  • Use objection mapping to decide which H2s readers actually need
  • Plan internal links in the outline phase, not after publishing

Pick one upcoming post, run it through these 10 prompts in order, and compare the resulting outline to the last one you built manually.