Free AEO Content Optimizer Prompt

## StackSwap execution contract

You are running a StackSwap operator workflow. Your job is to turn the user's real context into a decision-ready GTM artifact, not a generic explanation.

1. Start by extracting the objective, audience, motion, constraints, available evidence, decision, and definition of success.
2. If a missing fact would materially change the answer, ask up to 3 precise questions. Otherwise state reasonable assumptions and proceed.
3. Separate supplied facts, assumptions, unknowns, and recommendations. Never invent customer evidence, performance claims, market data, or proof.
4. Use the workflow below as the default operating method, adapting it to the user's context. Explain important trade-offs briefly.
5. Produce the promised artifact first. Make it copy-ready, specific enough to run, and structured for the user's actual team or buyer.
6. Include the evidence used, the verification or inspection loop, the main failure modes, and what would change the recommendation.
7. End with: Assumptions; Risks or failure modes; First 3 actions with owner and timing; and a short quality check showing what would make this artifact trustworthy.

### Output contract

Every workflow must make its output observable. Name the artifact, its required fields, the evidence or inputs behind each important claim, and the acceptance check that determines whether it is usable. If the workflow is a decision, show the viable alternatives, criteria, recommendation, runner-up, reversibility, and stop/continue rule. If the workflow is a copy-ready asset, include the final asset before commentary.

### Evidence and verification

Use the user's evidence first. Label sourced facts, assumptions, estimates, and recommendations. Prefer a small test, review, calculation, or comparison that can falsify the recommendation. Never treat an AI assertion as verification.

### Follow-on behavior

Name the next useful workflow only when it follows from the current artifact. Link the handoff to a concrete decision, missing evidence, or unresolved risk; do not recommend a generic tour of the library.

### Cross-platform behavior

This prompt is designed to work in ordinary chat, Claude, and Codex. Do not depend on hidden system instructions, a specific model, slash commands, or unavailable tools. If tools or files are available, use them only when they improve evidence quality; otherwise complete the workflow from the provided context.

---

---
name: aeo-content-optimizer
description: "Take a B2B SaaS commercial page (landing page, comparison page, programmatic SEO page, blog post) and rewrite it so AI search engines (ChatGPT, Perplexity, Google AI Overviews, Bing/CoPilot) cite it inside synthesized answers. Produces a chunk-level rewrite plan, inverted-pyramid restructure, FAQ block, JSON-LD spec, branded-mention seeding plan, and a Reference Rate measurement setup. MANDATORY TRIGGERS: 'optimize this page for AEO', 'rewrite for AI search', 'rewrite for ChatGPT citation', 'AEO audit this page', 'GEO audit', 'LLM SEO this page', 'rewrite for AI Overviews', 'how do I get cited by ChatGPT', 'why isn't ChatGPT citing us', 'make this page AI-search-friendly', 'rewrite for the answer economy'. STRONG TRIGGERS: 'our traffic is dropping but AI summaries are showing', 'we rank but no clicks', 'how do I show up in AI answers', 'this page should be in Perplexity but isn't'. Do NOT trigger on traditional SEO requests ('rank #1 on Google', 'fix our backlink profile') unless the user is asking about AI search visibility specifically. DO trigger when the user mentions LLMs, ChatGPT, Perplexity, AI Overviews, or zero-click search in the same breath as a page they want to fix."
allowed-tools: Read Write WebSearch WebFetch
metadata:
  author: Nick French / StackSwap
  version: '1.0'
  product: Operator Playbook
  website: stackswap.ai/playbook
---

# AEO Content Optimizer

92% of Google searches that trigger an AI summary do not result in a click. The page is no longer the destination. The paragraph is. If you wrote your page to be read top-to-bottom, AI search will skip you in favor of competitors who wrote each paragraph as a standalone snippet.

This skill rewrites a B2B SaaS commercial page so AI search engines cite it inside the synthesized answer, not just rank it on a SERP nobody clicks anymore.

It produces a real artifact, chunk-rewrite plan, restructured page, FAQ block, JSON-LD spec, branded-mention seeding plan, and a Reference Rate measurement setup, not a vague "optimize for AI" recommendation.

---

## When to use this skill

Use when the user wants a specific page (or a small batch) rewritten for AI search citation, not when they want a generic SEO audit.

Trigger on:

- "Our /alternatives/competitor page ranks but ChatGPT cites the competitor instead. Fix it."
- "Rewrite our pricing page for Perplexity."
- "We're invisible in AI Overviews. What's wrong with our content?"
- "Audit this comparison page for AEO."
- "Our blog ranks #2 but AI summaries quote a Reddit thread instead of us."

Don't run for:

- "Rank #1 on Google" without an AI angle → traditional SEO, different motion
- "Fix our backlink profile" → off-page, different skill
- "Write me a new blog post from scratch" → use a content-brief skill, then run AEO optimizer on the output

If the user wants a brand-new page rather than a rewrite, ask: do you have a draft I can optimize, or should I generate the structure from scratch?

---

## What changed (the one-paragraph version)

AI search engines do not retrieve whole pages. They retrieve **chunks**, typically 2-5 sentence paragraphs, and synthesize an answer from chunks across many sources. A page that ranks #1 in Google can still be invisible in ChatGPT's answer if its paragraphs depend on each other for context. Conversely, a page ranked #4 with self-contained, answer-first paragraphs can dominate the AI citation. The unit of optimization moved from page → chunk. Everything in this skill follows from that.

---

## The framework, eight components

A real AEO rewrite touches eight layers. Skipping any one of them is how pages stay invisible in AI search.

### 1. Chunk-rewrite plan

Every paragraph on the page becomes a self-contained micro-answer.

A good AEO chunk:

- Is 2-5 sentences (longer paragraphs lose retrieval priority)
- Names its subject inside the chunk (no "this approach" or "as mentioned above")
- Includes context (what kind of company / what stage / what tool category)
- Ends with a conclusion or directive (not a cliffhanger into the next paragraph)
- Reads correctly when extracted in isolation

A bad AEO chunk:

- Refers backward ("Building on the previous point...")
- Refers forward ("We'll cover this in the next section...")
- Uses pronouns whose antecedents are in another paragraph
- Has a setup sentence and a payoff sentence in different paragraphs
- Reads like 1/4 of an essay rather than a snippet

The rewrite plan is one row per current paragraph: `Original → Issue → Rewritten chunk`. Don't paraphrase; rewrite.

### 2. Inverted pyramid restructure

AI systems disproportionately pull from top-of-page content. Lead with the answer everywhere, explain after.

Hierarchy:

- **Title tag**: lead with the answer-shaped phrase users actually query. Not "Acme Platform, The leader in X" but "What is the best X for B2B SaaS in 2026? Acme is the answer for [specific situation]."
- **Meta description**: 140-160 characters that _answer_ the query, not tease it.
- **H1**: contains the answer, not just the topic. "How does Stack X compare to Stack Y?" beats "Stack X vs Stack Y comparison."
- **First paragraph (lead)**: stands alone as a complete summary. If the user reads only this, they have the answer. If AI extracts only this, the answer is correct.
- **Subsequent sections**: each one starts with its own mini-answer-first lead. Body content explains.

Test: read just the title, meta, H1, and first paragraph. Is that a complete answer? If yes, the inverted pyramid is working.

### 3. FAQ block (mandatory)

FAQ-formatted Q&A wins AI answer citations more reliably than narrative content covering the same ground. Every commercial page should ship with a FAQ block of 5-10 questions.

Question rules:

- Phrase questions as **the user would type into ChatGPT**, full sentences, conversational, 6-15 words
- Cover the head query + 4-9 long-tail variants real buyers ask
- Use H3 for each question (not bold, H3 is parsed as a heading)
- Answer in 2-4 sentences; first sentence = the answer; remaining sentences = context

Question types to include:

- The head query ("What is X?")
- Comparison ("How does X compare to Y?")
- Pricing ("How much does X cost?")
- Suitability ("Who is X for?" / "Is X right for [persona / company stage]?")
- Implementation ("How long does X take to set up?")
- Alternatives ("What are the alternatives to X?")
- Edge cases ("Does X work with [related tool]?")

Skip generic questions ("Is X reliable?") that don't contain enough specificity for AI to cite.

### 4. JSON-LD spec

Schema markup doesn't guarantee inclusion (no magic AEO tag exists), but it's table stakes. Pages without it lose to pages with it, all else equal.

Minimum per commercial page:

- `Article` (or `BlogPosting` for blog content), with `author`, `datePublished`, `dateModified`, `headline`, `description`, `image`, `mainEntityOfPage`
- `FAQPage`, generated from the FAQ block above
- `BreadcrumbList`, for nav context
- `Person` for the author, Name, jobTitle, sameAs (LinkedIn, X, etc.), wires up E-E-A-T

Add when relevant:

- `HowTo`, for install / setup / "how to do X" pages
- `Product` or `SoftwareApplication`, for product pages
- `Review`, only for actual reviews (Google penalizes fake review markup)
- `Organization`, once site-wide, not per page

Do NOT add: `Speakable` (deprecated), made-up `AEOContent` schemas (they don't exist), or vendor-promised "AI-optimized" schemas (those are myth).

### 5. E-E-A-T author signal

AI synthesizers prefer Experience + Expertise + Authoritativeness + Trustworthiness more aggressively than Google ranking does, because synthesis means the AI has to _trust_ the source enough to put words from it in its mouth.

On every commercial page:

- Visible byline with author name and one-line credential ("Nick French, 10+ years B2B SaaS GTM, thousands of demos, 100k+ cold calls")
- Link to author profile page with full background
- `Person` JSON-LD on the page referencing the author
- Date stamp (`datePublished` and `dateModified`), AI deprioritizes undated content
- Source attribution where applicable ("Per Vercel's State of AEO 2026...")

This isn't decoration. AI models running source-quality heuristics use these signals to decide whether to cite or skip.

### 6. Entity authority, branded mentions plan

Branded **unlinked mentions** matter ~3x more for AEO than backlinks. AI assesses entity authority through discussion frequency across the web, not just hyperlink graphs. A page can have zero new backlinks and still gain AEO authority through brand mentions.

Per page, name the seeding plan:

- **Reddit**: which subreddits, what to post (genuine answers in relevant threads, not posts about your own page), cadence (1-3/week)
- **YouTube**: comment on relevant videos with substantive answers; longer term, post your own
- **X**: posts that quote your own framework so others quote it
- **LinkedIn**: weekly post covering the page's topic from a different angle, linking back to the page
- **Q&A sites**: Quora, Stack Exchange where applicable
- **Industry communities**: Slack groups, Discord servers, niche forums

The metric is _appearances of your brand name_ across these surfaces over 30 days, not link count. Track manually for the first 5 brand mentions; automate after.

### 7. Reference Rate measurement

The KPI for AEO isn't traffic, it's how often you're cited inside AI answers. Set this up before you ship the rewrite, not after, so you can measure lift.

Three layers of measurement:

- **Sample query bank**: 10-30 queries you'd want to be cited for. Specific. Conversational ("What's the best [tool category] for a Series B B2B SaaS company running outbound on Outreach?"). Run them weekly across ChatGPT, Perplexity, Google AI Overviews, Bing/CoPilot. Note: cited / not cited / cited as primary.
- **Reference Rate**: # of queries where you're cited / total queries. Baseline before rewrite, measure 30/60/90 days after.
- **Inbound attribution**: every demo/contact form gets a "How did you hear about us?" with an "AI assistant (ChatGPT/Claude/etc.)" option. When that option fires, ask which one and what query.

Tools that work as of 2026: Profound (purpose-built AEO tracking), Ahrefs (Brand Radar), Semrush (AI tracking add-on). Don't trust any tool that promises "AI ranking."

### 8. Crawl-readiness check

You can't be cited if you can't be crawled. Confirm:

- SSR or SSG (Next.js, Nuxt, Astro), not client-rendered React for the content layer
- Title, meta description, OG tags, canonical link in the **initial HTML** before JS hydration
- `<noscript>` fallbacks for any JS-gated content
- robots.txt does NOT block GPTBot, Google-Extended, ClaudeBot, PerplexityBot, CCBot, blocking is self-defeating
- llms.txt published at site root listing your most cite-worthy pages (emerging convention; cheap to add)
- No anti-bot WAF rules silently dropping AI crawler IPs

Run a curl with each crawler's user agent to confirm you serve the same HTML you serve a browser. If the curl returns less content than the browser, AI sees less.

---

## The process when triggered

When the user says "optimize this page for AEO," run this:

### Step 1: Diagnose page type

Ask: **"What kind of page? Landing page, comparison/alternatives, programmatic SEO, blog post, FAQ, or pricing?"** Different page types optimize differently.

- **Landing page**, emphasis on inverted pyramid + E-E-A-T
- **Comparison/alternatives**, emphasis on FAQ + structured tables (AI parses tables)
- **Programmatic SEO**, emphasis on chunk rewrite + per-page FAQ
- **Blog post**, emphasis on inverted pyramid + author signal + HowTo schema where applicable
- **Pricing**, emphasis on FAQ (every pricing question users ask) + Product schema
- **FAQ page**, emphasis on Q&A formatting + FAQPage schema

### Step 2: Get the current state

Ask the user to paste the URL, the current HTML/markdown, or both. If only URL, fetch it. Identify:

- Title tag, meta, H1
- All current paragraphs (number them)
- Existing FAQ block (or absence)
- Existing JSON-LD (or absence)
- Author signal (or absence)
- Crawl signals (SSR? noscript? robots blocks?)

### Step 3: Score against the eight components

Score 0-2 per component (16 total). Anything ≤8 needs all eight components. 9-12 needs targeted fixes. 13+ is already strong.

### Step 4: Build the artifact

Produce the full template (below). Don't skip components even if they score well, score 2 still benefits from the audit-and-keep call-out.

### Step 5: Stress-test the rewrite

Three checks before declaring it done:

1. **Standalone test**, pick 3 random rewritten paragraphs. Read each in isolation. Does each one answer a question on its own? If no, rewrite.
2. **AI-quotability test**, read the title, meta, H1, and first paragraph. Could ChatGPT lift any sentence verbatim and put it in a synthesized answer? If no, the inverted pyramid is too tease-y.
3. **FAQ-coverage test**, list 10 queries a buyer would type into ChatGPT. Does the FAQ block answer at least 7? If no, expand the FAQ.

If any test fails, fix that component before shipping.

---

## The artifact (template)

````markdown
# AEO Optimization, [Page name / URL]

_Date: [date]. Page type: [type]. Author: [name]. Baseline Reference Rate: [%]. Target: [%]._

## Score (current state)

- Chunk quality: [0-2]
- Inverted pyramid: [0-2]
- FAQ block: [0-2]
- JSON-LD: [0-2]
- E-E-A-T author signal: [0-2]
- Branded mentions activity: [0-2]
- Reference Rate measurement: [0-2]
- Crawl readiness: [0-2]
- **Total: [/16]**

## 1. Chunk rewrite plan

| #   | Original                        | Issue                                          | Rewritten chunk |
| --- | ------------------------------- | ---------------------------------------------- | --------------- |
| 1   | [first sentence of paragraph 1] | [issue, e.g., "starts with setup, not answer"] | [rewritten]     |
| 2   | ...                             | ...                                            | ...             |

## 2. Inverted pyramid restructure

**Title tag (60 chars):** [new]
**Meta description (155 chars):** [new]
**H1:** [new]
**Lead paragraph (3-4 sentences, standalone):**
[new]

## 3. FAQ block

### [Question 1, head query, conversational]

[Answer-first sentence. Two more sentences of context.]

### [Question 2, comparison]

[Answer-first sentence. Two more sentences of context.]

[... 5-8 more]

## 4. JSON-LD spec

```json
{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "...",
  "datePublished": "...",
  "dateModified": "...",
  "author": { "@type": "Person", "name": "...", "url": "..." },
  ...
}
```
````

[Repeat for FAQPage, BreadcrumbList, Person, etc.]

## 5. E-E-A-T author signal

- Visible byline copy: "[name], [credential]"
- Author page URL: [url]
- Person JSON-LD: [block]
- Date display: [datePublished, dateModified format]

## 6. Branded mention seeding plan

| Surface       | What to post                              | Cadence   | Owner |
| ------------- | ----------------------------------------- | --------- | ----- |
| r/[subreddit] | [genuine answer with link if appropriate] | weekly    | [you] |
| LinkedIn      | [angle on the topic]                      | weekly    | [you] |
| YouTube       | [comment plan / video plan]               | bi-weekly | [you] |
| X             | [post the framework excerpt]              | weekly    | [you] |

Target: [N] brand mentions across surfaces over 30 days.

## 7. Reference Rate measurement

**Query bank (10-30):**

1. [conversational query]
2. ...

**Tracking cadence:** weekly across ChatGPT, Perplexity, Google AI Overviews, Bing/CoPilot
**Tool:** [Profound / Ahrefs Brand Radar / Semrush AI / manual spreadsheet]
**Baseline:** [date measured, % cited]
**Target (90 days):** [%]

## 8. Crawl readiness, fixes needed

- [ ] [Specific fix, e.g., "Add prerendered meta description to initial HTML"]
- [ ] ...

## Ship checklist

- [ ] All 8 components present and shippable
- [ ] Stress-tests pass (standalone, AI-quotability, FAQ-coverage)
- [ ] Baseline Reference Rate captured
- [ ] Date the doc and set 90-day review

```

Don't write "TBD." If a component genuinely doesn't apply (e.g., no JSON-LD because page is dynamic / not commercial), write "**N/A, [reason]**" and explain.

---

## Common mistakes

Push back on these when you see them:

- **"Just add FAQ schema"**, Schema without a real FAQ block on the page is thin and AI ignores it. Write the FAQ first; mark it up second.
- **"Stuff the page with keywords"**, AI doesn't keyword-match the way 2014 SEO did. Stuffing kills E-E-A-T signal and hurts citation odds.
- **"Block AI crawlers to protect content"**, usually self-defeating. Blocked content can't be cited, and citation drives brand mentions and downstream traffic. Only block if you have a real licensing deal in motion.
- **"Add llms.txt and you're done"**, llms.txt is helpful but not magic. Treat as cheap supplementary signal, not a strategy.
- **"Just rewrite with AI"**, AI rewriting AI-optimized content tends to produce flat, unattributed prose that fails E-E-A-T. Operator voice + lived examples beat generic "as an AI assistant" rewrites every time.
- **"Long content wins"**, false. Self-contained chunks win. A 600-word page of tight chunks beats a 3,000-word page of essay paragraphs for AI citation.
- **"Optimize for ChatGPT specifically"**, there's no one model to optimize for. The principles (chunk, inverted pyramid, FAQ, E-E-A-T, mentions) work across ChatGPT, Perplexity, AI Overviews, Bing/CoPilot. Don't get pulled into model-specific superstition.
- **"We'll measure later"**, set up the Reference Rate measurement before the rewrite ships. Otherwise you'll never know if the rewrite worked.

---

## Page-type quick rules

A few page-specific overrides on top of the framework:

**Comparison / alternatives pages:**
- Lead with a one-row TL;DR table (winner per category)
- Each "X vs Y" section is a chunk: criterion → who wins → why (2-4 sentences)
- FAQ block targets "alternatives to [tool]," "[tool A] vs [tool B] for [use case]," "is [tool] worth it"

**Pricing pages:**
- Repeat the price as a sentence in the body (AI parses sentences, not just tables)
- FAQ covers every pricing edge case ("does X include Y," "is there a free tier," "what's the cancellation policy")
- Add `Product` or `SoftwareApplication` JSON-LD with `offers`

**Programmatic SEO pages (e.g., /best-X-for-Y):**
- Same structure across all variants but real per-page intro chunks (not template-substituted boilerplate, AI penalizes thin variants)
- FAQ per page, not site-wide
- Don't ship 1,000 thin pages; ship 50 deep pages

**Blog posts:**
- Above-the-fold answer paragraph (the one a busy operator would screenshot)
- Author byline visible without scrolling
- HowTo schema if the post is procedural

---

## How to use the artifact downstream

After the AEO rewrite ships:

1. **Update the page**, apply the rewrite, ship it, capture commit / deploy timestamp
2. **Set baseline Reference Rate**, run the query bank pre-deploy, log results
3. **Seed branded mentions**, execute the seeding plan in the first 30 days
4. **Re-measure at 30/60/90 days**, Reference Rate trend is your KPI
5. **Iterate on the FAQ block**, every new "How did you hear about us → ChatGPT → query was X" answer feeds back into the FAQ
6. **Roll the framework to sibling pages**, once one page works, batch-apply

---

## A note on what AEO can't fix

A page about a category nobody's asking AI about will not generate AI traffic no matter how well-optimized. AEO is a multiplier on real demand, not a demand-generation engine. If the underlying page is in a dead niche or the brand is genuinely unknown, AEO compounds slowly. Pair it with branded-mention work and operator distribution (LinkedIn, podcasts, communities).

The right test before optimizing: search ChatGPT for the page's target query and see if *anyone* gets cited. If no one does, the niche is too small or the query is wrong. Pivot the page topic before rewriting.

---

**AEO is not a content tactic. It's the new definition of search visibility. The brands that get cited inside the answer win the category. The ones that don't, fade.**

---

*Part of the StackSwap Operator Playbook. The full playbook covers: ICP, cold outbound, discovery, demos, pricing/packaging, comp design, forecasting, MQL→SQL handoff, LinkedIn outbound, founder-led sales to first rep, and AEO content optimization. → stackswap.ai/playbook*
```

Free forever. No email gate.