Get Visibility in ChatGPT, Perplexity, Google AI Search in 2025 (With Prompts) - AEO / GEO Playbook - Part 2 of 3
Optimize your content for 2025 AI Search visibility using tactics and prompts! This playbook (Part 2) guides marketers in AEO / GEO, ensuring your content appears in AI Overviews & answers.
Since last week’s newsletter on Part 1 of this series “Playbook - Transform SEO Content to ChatGPT, Perplexity, Google AI Optimized Assets for AEO / GEO in 2025 (With Prompts)”, I was somehow exposed to more online discussions around AEO / GEO…
Is this because I have written about it ? Or Is this because the topic is BUZZ”?
Before we start... If you are reading for the first time.
SEO - Search Engine Optimization (You already might have heard of it quite often)
The buzz words are AEO and GEO
1. Answer Engine Optimization (AEO) – optimizing content to be the answer delivered by search engines and voice queries.
2. Generative Engine Optimization (GEO) – optimizing for AI-driven search results (like Google AI Overviews, Bing Chat, ChatGPT, Perplexity etc.).
In case you have missed my last newsletter, which was also the first part of this series, here is the link...
Playbook - Transform SEO Content to ChatGPT, Perplexity, Google AI Optimized Assets for AEO / GEO in 2025 (With Prompts) - Part 1 of 3
Search is evolving from a “list of links” into “personalized answers”.
I saw this post by Nathan Latka…
“I got 31,937 clicks from ChatGPT and Perplexity
Here's how:
I have no idea.
My guess:
Content on GetLatka is unique CEO interview content
Its getting cited in LLM responses
Our investments in traditional SEO are helping”
What is he doing differently? - “Talking about what people want and search for… in a language meant for humans rather than Google Search”
Here is another research by “Profound”:
What it means
For ChatGPT:
Structure wins. Use clear headings, schema markup, and tight step-by-step blocks so the model can refer to your answer.
For Google AI Overviews:
Go social. Fresh opinions from Reddit, YouTube, and niche forums are preferred over just articles. Community insight gets the spotlight.
Here is what happening on “Reddit” - Link
Everyone is saying the same thing… “Game is changing”.
And Google also announced “AI in Search: Going beyond information to intelligence”
And this is what they mentioned…
“Under the hood, AI Mode uses our query fan-out technique, breaking down your question into subtopics and issuing a multitude of queries simultaneously on your behalf. This enables Search to dive deeper into the web than a traditional search on Google, helping you discover even more of what the web has to offer and find incredible, hyper-relevant content that matches your question.”
We have covered the following Phases in Part 1:
Phase 1: Audit Your Content and Current Search Performance - From SEO to AEO / GEO Visibility in 2025
— AI Prompt - Audit a Page for Answer Readiness
Phase 2: Prioritize and Plan Your AEO (Answer Engine Optimization) / GEO (Generative Engine Optimization) Updates
— AI Prompt - Prioritize Content Topics
Phase 3: Rewrite and Optimize Content for AEO / GEO (using AI Prompts)
— AI Prompt - Rewrite for a Featured Snippet
— AI Prompt - Generate Q&A from Content
Now let’s cover the remaining phases which we left in the last part.
Phase 4: Enhance with Schema Markup and Technical SEO Signals
Phase 5: Monitor, Measure, and Iterate for Continuous Improvement
Want more playbooks like this?
→ Follow me on LinkedIn
→ Subscribe to StartupGTM
→ Bookmark this post and share it with your team
If you’re here for the first time, below is a list of most liked articles:
The Lenny Rachitsky Playbook : Prompts, Growth Frameworks, and Strategies - Part 1 of 2
AI (ChatGPT, Gemini, Claude) Prompts for CMOs, Marketers and Growth Builders - Part 1 of 3
Below are some important prompts that would be useful while you are enabling your content for AEO / GEO…
Prompt #1 ( Specific to AEO ) - List of long‑tail, conversational queries and semantic terms for [YOUR TOPIC]
As a Chief Marketing Officer, you want ChatGPT to act as your strategic SEO partner. Here’s a turnkey prompt you can drop into ChatGPT (or any LLM), replacing [YOUR TOPIC] with your chosen niche:
Role & Task:
“You are a seasoned Chief Marketing Officer with 20 years of B2B marketing experience. Your goal is to generate highly actionable, SEO-driven content ideas.Instructions:
1. Generate 30–50 long-tail, conversational search queries for [YOUR TOPIC].
2. Categorize each query by search intent:
Informational (e.g., “how to…”, “what is…”)
Navigational (e.g., “best [YOUR TOPIC] platforms”, “[Brand X] pricing”)
Transactional (e.g., “buy [YOUR TOPIC] online”, “[YOUR TOPIC] discount code”)
3. For each query, extract and list the key semantic entities or modifiers (e.g., “free trial”, “comparison”, “reviews”, “near me”, “use cases”).
4. Ensure an even distribution of queries across TOFU, MOFU, and BOFU buyer-journey stages. Include at least five “question-style” phrases per intent.
5. Format the output as a JSON array of objects, for easy ingestion into content planning tools, following this structure:[
{
"query": "how to choose the best [YOUR TOPIC] solution for startups",
"intent": "informational",
"entities": ["best", "solution", "startups"],
"funnel_stage": "TOFU"
},
{
"query": "pricing comparison of top [YOUR TOPIC] tools",
"intent": "navigational",
"entities": ["pricing", "comparison", "tools"],
"funnel_stage": "MOFU"
},
{
"query": "buy enterprise-grade [YOUR TOPIC] software discount",
"intent": "transactional",
"entities": ["buy", "enterprise-grade", "discount"],
"funnel_stage": "BOFU"
}
// …more entries…
]
Tone & Style:
Friendly yet authoritative
Conversational, reflecting real voice-search patterns
Data-driven: prioritize queries with clear buyer intent or problem-solving focusDeliverable:
A JSON array of 30–50 objects, evenly split across informational, navigational, and transactional intents, each with its query, intents, semantic entities, and funnel stage.”
Prompt #2 ( Specific to AEO ) - Create FAQ entries: ‘What is X?’
You are a seasoned Chief Marketing Officer with 20 years of B2B SaaS marketing leadership.
Your mission is to produce high-impact FAQ entries that engage prospective buyers and rank for featured snippets.
Instructions:
1. Question Format
- Focus on “What is [X]?” as the FAQ question.
- Render it as an H3 Markdown header: `### What is [X]?`
2. Answer Requirements
- Provide one concise paragraph of exactly 40–50 words.
- Write in a friendly yet authoritative tone, optimized for scannability and trust.
3. Key Facts/Stats
- Follow the paragraph with a bulleted list of 3–5 succinct facts or statistics.
- Each bullet should reinforce credibility (e.g., user counts, performance improvements, industry benchmarks).
4. SEO & Brand Alignment
- Weave in the core value proposition of [X] where natural.
- Prioritize metrics or social proof (e.g., “trusted by 10,000+ companies,” “reduces costs by 20%”).
5. Output Example
```markdown
### What is [X]?
[40–50-word answer here.]
- Used by over 10,000+ mid-market companies worldwide
- Automates reporting to reduce manual work by 30%
- Integrates with 50+ enterprise tools out of the box
Deliverable:
A set of Markdown-formatted FAQ entries for “[X],” each following the above structure, ready to publish on your website or feed into your chatbot.
Prompt #3 - “People Also Ask Questions related to [TOPIC]
You are an expert content strategist fluent in SEO, UX writing, and knowledge-base development. Your task is to mine high-value user questions and deliver concise, answer-first content with structured support.
Instructions:
1. Extract PAA Questions
- For “[TOPIC]”, retrieve the top 15–20 “People Also Ask” questions from Google.
- Group them by search intent: informational, navigational, or transactional.
2. Answer-First Paragraph
- For each question, write a 1–2 sentence answer that immediately addresses the query.
- Start with the key fact or definition (e.g., “X is…”) before elaborating.
3. Supporting Structure
- Under each answer, include 2–3 subheaders (H3 or H4) that outline deeper points or common follow-ups.
- Beneath each subheader, add 1–2 bullet points of evidence—statistics, examples, or brief citations.
4. Formatting
- Use Markdown.
- Clearly label each intent section (e.g., `## Informational Intent`).
- For each question:
```markdown
### Q: [User question here]
A: [Direct answer-first text, 1–2 sentences.]
#### [Subheader 1]
- [Evidence bullet]
#### [Subheader 2]
- [Evidence bullet]
```
5. Tone & Style
- Authoritative yet approachable—suitable for both technical and business audiences.
- Optimized for featured snippets and quick scanning.
Deliverable:
A Markdown document containing all PAA questions for “[TOPIC],” organized by intent, each with a direct answer followed by structured headers and evidence.
Prompt #4 - Analyze competitor AI‑generated answers for prompt ‘top-rated [PRODUCT TYPE]’
You are a seasoned Chief Marketing Officer, AI technologies expert, and content strategist with 20 years of experience driving data-led growth. Your mission is to benchmark and out-perform competitor AI-generated content on the query “top-rated [PRODUCT TYPE]” by matching their structural rigor and citation depth, then crafting an optimized outline that exceeds their standard.
Instructions:
1. Competitor Retrieval (AI-Generated Answers)
- Query “top-rated [PRODUCT TYPE]” using three leading AI content tools or platforms.
- Collect the top 3–5 AI-generated responses, ensuring they each include headings and citations.
2. Structural Analysis
- For each response, extract its H2/H3 (and deeper) heading hierarchy.
- Note the logical flow and section lengths (e.g., introduction, feature breakdown, comparisons, conclusion).
3. Citation Benchmarking
- Tally the number of citations per section and categorize by type:
- Data/Stats (e.g., “Forrester report, 2024”)
- Expert Quotes (e.g., “Jane Doe, CTO at X”)
- Case Studies/Examples
- External Links
- Compute average citations per 500 words and identify any sections with unusually high or low citation density.
4. Competitive Gap & Opportunity Identification
- Highlight sections where competitors:
- Lack sufficient evidence or examples
- Skip relevant subtopics (e.g., total cost of ownership, integration steps)
- Use generic headings that could be more benefit-driven
5. Outline Creation to Match & Exceed
- Produce a Markdown outline for your “top-rated [PRODUCT TYPE]” article, including:
- Matched or improved H2/H3 structure
- New subtopics to fill identified gaps (e.g., “Real-world ROI Case Study,” “Comparison Matrix: 5 Leading Models”)
- Recommended citation targets per section (e.g., “3 data cites + 1 expert quote”)
- Calls to action or interactive elements (e.g., “embedded comparison table,” “customer testimonial video”)
6. Format & Deliverable
- Part A: Side-by-side table listing each competitor’s headings and citation counts
- Part B: Your enhanced Markdown outline with clear guidance on structure and citation frequency
- Part C: Bullet-pointed action plan under each heading, specifying how to surpass competitors (e.g., “Add original survey data here,” “Include a 200-word executive quote”)
Tone & Style:
- Strategic yet tactical—suitable for board-level decision-makers and hands-on content teams
- Data-driven, with clear metrics and actionable recommendations
Prompt #5 ( Specific to GEO ) - Draft an explainer post and test via Perplexity/OpenAI and refine structure.
You are a veteran Chief Marketing Officer, AI expert, and content strategist with 20 years of experience translating complex topics into engaging, beginner-friendly content. Your task is to draft and iterate an explainer post titled “GEO for Beginners.”
Part A: Draft
1. Title & Introduction
- Title: “GEO for Beginners”
- Write a 2–3 sentence intro that hooks a reader who’s never heard of GEO, using a conversational, encouraging tone.
2. Question-Based Headings
- Create 5–7 H2 headings, each framed as a learner’s question (e.g., “What Is GEO?” “Why Does GEO Matter?”).
- Under each heading, write 150–200 words at a grade-6 reading level (use short sentences of 8–12 words, simple vocabulary).
3. Style & Tone
- Conversational and friendly, as if explaining to a curious middle-schooler.
- Avoid jargon; when a technical term is unavoidable, define it briefly in plain language.
Part B: Testing & Refinement
4. Automated Clarity Check
- Run your draft through an AI text-analysis tool (e.g., Perplexity.ai or OpenAI’s own evaluation API) to measure:
- Perplexity: target ≤ 20
- Burstiness: target low variation for predictability
- Readability Score: confirm grade-6 level (Flesch-Kincaid or similar)
5. Structural Refinement
- Based on the tool’s feedback, identify:
- Any headings or sections that scored above target perplexity or reading-level thresholds.
- Sentences flagged as too complex or scores indicating uneven pacing.
- Revise by:
- Splitting long sentences, substituting simpler words, or reordering sections for flow.
- Tweaking headings for clarity and learner curiosity.
6. Final Deliverable
- Section 1: The refined Markdown draft with title, intro, question-based headings, and body text.
- Section 2: A brief “Refinement Log” summarizing the automated scores pre- and post-edit and listing the key structural or stylistic changes made.
Prompt #6 - List 5 strategies to restructure [YOUR PAGE] for better LLM extractability
You are a seasoned Chief Marketing Officer, AI technologies expert, and content strategist with 20 years of experience optimizing content for both human readers and large language models. Your mission is to recommend five concrete strategies to restructure “[YOUR PAGE]” so that LLMs can extract and repurpose its content more effectively.
Instructions:
1. Number of Strategies
- Provide exactly 5 distinct strategies, each focused on a different structural or formatting approach.
2. Description Requirements
- For each strategy, write 2–3 concise sentences explaining why it improves LLM extractability and how to implement it.
3. HTML Element & Example
- Specify the primary HTML element involved (e.g., `<table>`, `<ul>`, `<h2>`) for each strategy.
- Include a brief example snippet (1–2 lines of Markdown or HTML) illustrating the approach.
4. Tone & Style
- Conversational yet authoritative—speak as both a marketer and a technical advisor.
- Action-oriented: focus on clear, implementable steps.
5. Output Format
- Render your answer as a Markdown numbered list.
- Example format for one strategy:
```markdown
1. Use Descriptive Tables
- HTML element: `<table>`
- Example:
```html
<table>
<tr><th>Feature</th><th>Benefit</th></tr>
<tr><td>Auto-summarize</td><td>Reduces reading time</td></tr>
</table>
```
```
Deliverable:
A Markdown-formatted list of 5 strategies, each with its description, HTML element, and example snippet, ready for your development and content teams to implement on “[YOUR PAGE].”
Prompt #7 - Build llms.txt or metadata file listing your key entities/topics to guide crawlers and LLM attribution / visibility
You are a seasoned Chief Marketing Officer, AI technologies expert, and content strategist with 20 years of experience aligning marketing, engineering, and AI teams. Your task is to generate a machine-readable metadata file (e.g., `llms.txt`, JSON, or YAML) that captures your organization’s key entities/topics and their canonical forms, ensuring consistent crawler indexing and LLM attribution.
Instructions:
1. Choose Format
- Select one: plain text (`llms.txt`), JSON, or YAML.
- Ensure the file is UTF-8 encoded and easily parseable by both web crawlers and LLM ingestion pipelines.
2. Define Entity Schema
For each core entity/topic, include the following fields:
- `entity` (string): canonical name
- `aliases` (array/string list): common variants or synonyms
- `type` (string): category, e.g., `product`, `service`, `person`, `concept`
- `priority` (string): `high` / `medium` / `low` relevance
- `description` (string, optional): one-sentence definition or role
3. Populate Core Entities
- List all major products, services, brands, people, and domain concepts your organization owns or emphasizes.
- Normalize spelling variations, abbreviations, and alternative phrasings under each `aliases` array.
4. Example Entries
- YAML
```yaml
- entity: Acme AI Analytics
aliases:
- Acme Analytics
- Acme Analytx
- Acme AI
type: product
priority: high
description: "A unified AI-driven analytics platform for real-time business insights."
- entity: Jane Doe
aliases:
- J. Doe
- Jane D.
type: person
priority: medium
```
- JSON
```jsonc
[
{
"entity": "Acme AI Analytics",
"aliases": ["Acme Analytics","Acme Analytx","Acme AI"],
"type": "product",
"priority": "high",
"description": "A unified AI-driven analytics platform for real-time business insights."
},
{
"entity": "Jane Doe",
"aliases": ["J. Doe","Jane D."],
"type": "person",
"priority": "medium"
}
]
```
5. Deliverable
- A single metadata file named `llms.txt` (or `.json` / `.yaml`) containing your complete entity list.
- Ensure consistent indentation, valid syntax, and UTF-8 encoding.
Now, continue with the remaining phases (it's high time in 2025)…
Phase 4: Enhance with Schema Markup and Technical SEO Signals (AI Search / AI Overviews look for these)
With content optimized in plaintext as per Phase 2 and Phase 3, we now add the structured data and technical enhancements that signal answer-readiness to search engines and AI.
Schema markup in particular is a critical bridge between your content and AI understanding.
1. How Do You Add Schema Markup for AI Search?
Schema.org markup helps search engines interpret content context, which in turn helps AI models use your content correctly.
By adding JSON-LD or microdata to your pages, you can explicitly indicate questions, answers, how-to steps, product details, and more.
Here are some high-value schema types for AEO/GEO and how they help:
Schema Markup Type #1: FAQPage
Mark up a list of questions and answers on a page.
How It Helps AI/SEO?
Allows Google to feature your Q&A directly in search results (rich FAQ snippets), and helps AI identify common Q&A to include in answers. Great for pages with an FAQ section.
Schema Markup Type #2: HowTo
Mark up step-by-step instructions (with steps, tools, durations, etc.).
How It Helps AI/SEO?
Enables rich how-to results with visual steps. AI can easily extract the sequence of steps for procedural queries (“How to…”).
Schema Markup Type #3: Article/BlogPosting
Mark general content with headline, author, date, etc.
How It Helps AI/SEO?
Ensures the AI knows this is an article. Include author (with credentials if possible) to support E-E-A-T – authoritative content is more likely trusted by generative AI.
Schema Markup Type #4: Product
Mark product details (name, description, SKU, brand, etc.) and nested Offer, AggregateRating, Review where applicable.
How It Helps AI/SEO?
Powers rich results for product queries (price, availability, star ratings in SERPs). Also helps AI understand your product in context. If a generative AI is asked “What’s the best [product]?”, it might utilize schema info like reviews or attributes in its comparison.
Schema Markup Type #5: Review / AggregateRating
Mark customer reviews and overall rating.
How It Helps AI/SEO?
Displays stars and ratings in Google results, and gives AI a sense of user sentiment. Generative answers might mention “rated 4.5 stars based on 200 reviews” if your schema provides it.
Schema Markup Type #5: Organization / LocalBusiness
Mark up your business information (industry, location, contact, etc.).
How It Helps AI/SEO?
Feeds Google’s Knowledge Graph and local panels. Critical for brand queries and local voice searches (“Find a __ near me”). Also ensures AI has accurate facts on your company (e.g., founding date, CEO) .
Schema Markup Type #6: Speakable (beta)
Mark a section of the page best suited for text-to-speech.
How It Helps AI/SEO?
For news or articles, this indicates a snippet of text for voice assistants to read aloud. If your content is often accessed via voice (e.g., newsy blog or Q&A), marking a 2-3 sentence summary as speakable could increase chances Google Assistant uses your content.
Schema Markup Type #7: ImageObject / VideoObject
Mark up images and videos with descriptions.
How It Helps AI/SEO?
Helps search engines (and thus AI) interpret visual assets. For example, if a how-to has a video, marking it up might let an AI search include a video thumbnail or at least know there’s a visual available.
Implementing schema can be done regardless of CMS. If you use WordPress or another CMS, plugins like Yoast or RankMath can simplify adding FAQ, HowTo, Article markup, etc..
In custom setups, you might need to inject JSON-LD in the HTML. Google’s Structured Data Markup Helper is a handy tool: you can paste your page and visually tag content to generate JSON-LD. Once added, use the Rich Results Test to validate (catch any errors).
2. Leverage Other “AI Signals”:
Beyond schema, there are subtle technical and on-page signals that can improve AI visibility:
Meta Tags for Context:
Ensure your <title> and meta <description> are concise and include the question or keywords the page answers. The meta description might not directly boost ranking, but it often becomes the snippet text on SERPs.
More importantly, a well-crafted meta that summarizes the answer could be used by generative AI as a quick reference.
For instance, if the meta description of an FAQ page says “Answers to common questions about [Topic], including how it works, pricing, and best practices.” – an AI might pick up phrasing from it when summarizing your site’s content.
Also consider the meta og: tags and twitter: cards – while for social sharing, sometimes these concise summaries help any context extraction tool. (No need to stuff keywords – focus on clearly stating what the page offers).
Heading Structure (HTML5):
Use proper heading levels (<h1>…<h2>…<h3>) to reflect the hierarchy of questions and answers. This structural HTML is an implicit schema that AI uses to understand content sections.
A clear outline of topics and subtopics in headings can improve how snippets or AI break down your content. For example, SGE might show separate bullet points for different sub-sections of your article if the headings are clear.
Clean HTML and Accessibility:
Remove any unnecessary clutter that might confuse parsers (like excessive ads, or non-semantic layout). Use lists for lists, tables for tabular data.
Ensure images have descriptive alt text – not only for accessibility, but if an AI can’t find text answer, it might look at alt text of an infographic or such.
If you have important info in images (like an infographic chart with stats), also describe that stat in text nearby so it’s not missed.
Page Speed and Mobile:
While this is SEO 101, it’s worth noting: if your content is slow or poorly loading on mobile, it might impact its crawl frequency or how users interact (e.g., if an SGE result includes a “click for more” from your site and your page is slow, the user may abandon – negative engagement).
Google’s algorithms, including for SGE, likely consider core web vitals to some extent. So optimize performance and mobile responsiveness.
Internal Linking & Navigation:
Add links to and from your newly optimized content.
For example, if you created a new FAQ page or a guide, link to it from relevant higher-level pages (“Learn more about how it works in our FAQ”).
A voice search might not use that link, but Google’s discovery of your content improves with strong internal linking. It also helps distribute any link equity, indirectly boosting rankings which as noted correlates with AI visibility.
Sitemap and Indexing Signals:
If you substantially updated a page for AEO, you might want to ping Google for re-indexing (use Search Console’s URL inspection -> Request indexing).
Also ensure your XML sitemap is updated with any new pages (like new FAQs). This gets the content into the index faster, which is important for AI – particularly if you included timely information or new data that you want the AI to pick up.
Monitor Robot Instructions:
Generally, you want your content accessible. Don’t use noindex or noarchive on pages you want featured in answers.
Conversely, if there are pages you don’t want used by AI (maybe because they are low quality or meant for other purposes), consider robots rules or the newer data-nosnippet (to prevent snippet usage of certain parts). But use carefully; our aim is mostly to open up content to be shared.
3. Optimize Off-Page Data for Brand:
While not exactly on-site technical, a quick note: ensure your Knowledge Graph and external sources about your brand are robust.
AIs often draw on sources like Wikipedia, Wikidata, Google My Business, etc. for factual queries about organizations.
If applicable, try to get a Wikipedia entry for your brand or key products (with neutral, factual tone). Keep your Google Business Profile updated (especially for local businesses).
This is part of sending the right signals that your brand is an authority and trustworthy.
Checklist – Technical & Schema Phase:
[✓] Add schema relevant to each page (FAQ, HowTo, Product, etc.) and validate it.
[✓] Mark organization details (Organization/LocalBusiness schema, Knowledge Graph entries) for brand presence.
[✓] Use proper HTML structure: headings, lists, table markup – no important info hidden in non-text formats.
[✓] Update meta tags with concise summaries including the question/answer focus.
[✓] Ensure crawlability & speed: fast load, mobile-friendly, and no robots.txt or meta tags blocking important content.
[✓] Allow AI access: Unless you have privacy concerns, permit known AI bots (e.g., don’t disallow GPTBot if you want ChatGPT to train on your site).
[✓] Re-index and test: Once changes are live, use Search Console to fetch and index. Test how your rich results appear (use the Rich Results tester and do some manual Google queries).
AI Prompts – Schema & Technical Aids:
You can even employ AI to help generate some of the structured data or identify technical issues. For example:
AI Prompt to Generate FAQ Schema:
You are an expert in SEO, AI search optimization, and structured data implementation.
Generate a valid and fully structured JSON-LD schema using the FAQPage format from [Schema.org], based on the following list of questions and answers about [insert topic]:
Q1: [insert question 1]
A1: [insert answer 1]Q2: [insert question 2]
A2: [insert answer 2]Q3: [insert question 3]
A3: [insert answer 3]Instructions:
— Output only valid, clean JSON-LD (no commentary or explanation).
— Use "@context": "https://schema.org" and "@type": "FAQPage".
— Wrap each Q&A in "@type": "Question" and "@type": "Answer" respectively
— Escape all special characters correctly to avoid JSON errors.
— Ensure the output complies with Google Rich Result guidelines for FAQPage.
Optional flags you may set before output (if needed):
— "language": to specify multilingual JSON-LD output (e.g., "en", "es")
— "pretty": for readable formatting instead of minified JSON
This can save time writing schema by hand (just be sure to verify the output). For checking content structure, you might ask:
AI Prompt for Technical Audit Check:
I will provide the HTML content of a web page.
Analyze the structure with a focus on clarity, semantic markup, and content hierarchy — specifically the use of headings (<h1>–<h6>), lists (<ul>, <ol>), and relevant HTML5 tags (e.g., <article>, <section>, <main>).Identify opportunities to improve:
-- Content clarity and accessibility for both users and search engines
-- Schema.org structured data implementation (e.g., FAQPage, Article, Product)
-- Semantic HTML enhancements to better signal content intent to AI systems and search crawlersReturn a breakdown of key issues and practical recommendations for improvement. Focus on SEO, featured snippet eligibility, and AI-based content understanding.
The AI won’t catch everything a dev or SEO pro would, but it might point out, for example, “It looks like there’s no FAQ schema even though there’s a Q&A section,” or “The headings skip from H2 to H4 – consider nesting properly.”
Phase 5: How do you Monitor, Measure, and Iterate for Continuous Improvement? (Crucial element of Playbook for AI Visibility)
Optimizing for AEO/GEO isn’t a one time project.
It requires ongoing monitoring and tweaking. In this phase, we focus on measuring results and using them to refine the strategy.
1. Track Key Metrics :
Define metrics that indicate success in answer optimization. Some important ones to monitor:
Featured Snippet Count:
How many snippets your site owns (and for which queries). You can track this via SEO tools or even by monitoring Search Console for queries where your average position is 1 and the impressions vs clicks suggest a snippet (low clicks but high position). There are also third-party tools that specifically track featured snippet presence.
People Also Ask Presence:
Monitor if your content is appearing as answers in PAA boxes. This might be harder to quantify, but keep an eye on common questions and see if your site shows up.
Voice Search Traffic:
If you have analytics that break out voice (some analytics can show queries from voice assistants, or you might infer from query phrasing), see if it increases.
Also, check Google Search Console’s Performance > Search Appearance for any filter like “Rich results” or “Discover” that might hint at appearances in new contexts (Google may eventually report SGE impressions here).
AI Referral Traffic:
Look at your analytics for referrers like gemini.google.com, chatgpt.com, search.app.goo.gl (SGE), bing.com/chat, or any spike in direct traffic after an AI rollout.
As an example, after implementing AEO changes, an SEO agency’s client saw 2,300% growth in monthly AI-driven traffic from platforms like ChatGPT and Google’s SGE.
AI referral traffic is emerging as a metric – tools and analytics definitions are still catching up, but any uptick is a good sign.
Brand Search and Mentions:
Monitor if searches for your brand name increase (indicating more people saw your brand in an answer and got curious).
Also, manually do some generative queries in AI chatbots related to your domain and see if your brand or content is mentioned/cited. For instance, if you ask ChatGPT or Bing, “What’s the best solution for X?” – does it mention your company now? Keeping tabs on this provides qualitative evidence of brand visibility.
Engagement & Conversion Metrics:
Paradoxically, AEO/GEO success might mean fewer clicks for certain queries (because users got their answer). So traditional traffic might not tell the full story.
Complement it with engagement metrics on the site for those who do click (time on page, scroll depth – are they finding value quickly?). Also track conversions that might stem from zero-click touchpoints: for example, a user heard your brand via voice and later did a direct visit or store visit.
This is hard to attribute, but any overall uplift in conversions or branded traffic over time can be partly credited to improved answer visibility (especially if you see brand search rising).
2. Collect Feedback and Observe AI Outputs:
Listen to what your customers or community are saying.
Do they mention “I found you through Google’s answer” or “Assistant told me about your product”? Those anecdotes are gold. Additionally, periodically use various AI platforms as a user: Ask the questions you’ve optimized for and see how the answers appear.
Does the AI pick up your content? Is it attributing correctly? For example, Google SGE typically cites sources with a hyperlink in the overview – check if your page is listed there when it should be. If not, analyze why the AI might be favoring another source. Perhaps the competitor had a more straightforward answer or a statistic that you could incorporate next update.
3. Fine-Tune Based on Performance:
Use the data to refine your approach:
If certain pages still aren’t getting the snippet, try tweaking the answer length or phrasing. Sometimes moving from a paragraph to a bullet list can do it. Or adding a more specific question header.
If an AI summary pulled text from your page but without mentioning your brand (maybe because the text snippet it grabbed didn’t include it), consider editing that sentence to include your brand next to the fact or claim.
Example: instead of “10% improvement in Y,” say “A 10% improvement in Y (as per [YourCompany]’s analysis)”. Subtle change, but now any quote of that text carries your name.
If you notice a particular new question trending (e.g., through Google’s “People Also Ask” or from customer inquiries), update your content or add a new FAQ to address it. Being fast and first with a good answer can secure you as the featured snippet or AI reference before competitors.
4. Scale Up and Institutionalize:
After initial successes, roll out the content transformation to the wider library in phases.
Create internal guidelines for writers so all new content is by default AEO-optimized (e.g., a checklist: did you phrase the heading as a question? Did you answer it in 1-2 sentences immediately? etc.).
Make AEO/GEO part of your content SOP. Similarly, for technical teams, ensure schema and structured data are part of the development checklist for new pages.
Essentially, bake this into your content operations so you don’t slip back into old SEO-only habits.
5. Stay Informed on AI Search Updates:
The landscape is rapidly changing. Google might alter how SGE presents sources; new AI search engines may emerge (e.g., OpenAI’s web search or others).
Keep an eye on industry news and adapt.
For example, if Google launches a way to opt-in to better AI visibility (they haven’t yet beyond schema, but one never knows), be ready to implement it.
Likewise, if studies (like the GEO benchmark) reveal new factors (suppose they find that “including an image with infographic data increases AI usage”), consider testing that. AEO/GEO is a new frontier – those who experiment and iterate will lead.
Case in Point – Iteration: One study introduced the term GEO and found that using certain strategies (like adding citations and relevant quotes) boosted content visibility in AI results by up to 40%.
But it also noted the effectiveness varies by domain. The lesson is to iterate and find what works in your niche. You may discover, for example, that definition boxes work well for your industry queries, or that adding a comparison table draws in AI answers for “A vs B” queries. Use your monitoring to drive these micro-optimizations.
AI Prompt Template – Ongoing Tuning:
AI can assist in monitoring and ideation too. For example:
AI Prompt to Explain AI Behavior:
You are an expert in SEO, AI search optimization (including Google SGE), and content structuring for machine-readable formats.
I queried Google’s AI-generated search experience (SGE) using:
‘<your query>’,
but my site did not appear in the AI-generated summary, despite having highly relevant content.Below is a snippet of that content:
<insert content snippet>Instructions:
— Analyze potential reasons why this content was not chosen by SGE.
— Evaluate content relevance, depth, semantic clarity, and topical alignment with the query.
— Assess HTML structure, use of schema.org markup, and on-page SEO signals.
— Identify any gaps in trust, authority, or freshness that might influence ranking in generative AI responses.
Deliverable:
Provide a prioritized list of 5–7 actionable recommendations to improve the content’s visibility and eligibility for inclusion in AI-generated answer panels — especially those used in Google SGE and similar AI search engines.
This can give you some creative ideas (it might respond with “Your content might not have the exact question phrased, or perhaps it’s too long-winded; try adding a concise answer…” etc., echoing principles you’ve learned).
AI Prompt to get Content Refresh Ideas:
“Given the following answer content I have for [question], suggest any additional information or updates that could make it more comprehensive or trustworthy for 2025. Content: <your answer>.”
The AI might suggest adding a recent statistic or addressing a sub-question, which you can then verify and include, keeping your answer content fresh and competitive.
Which Brands Successfully Rank in ChatGPT Search? AEO (Answer Engine Optimization) / GEO (Generative Engine Optimization) Success
To solidify the benefits of this transformation, let’s look at a few examples of organizations that embraced AEO/GEO tactics:
Go Fish Digital: Direct ChatGPT Search Manipulation
Company: Go Fish Digital (Digital PR Agency)
Website: gofishdigital.com
Challenge: Competitors were getting "Notable Clients" listings in ChatGPT Search results while Go Fish Digital wasn't
Timeline: 1 week implementation period
The Strategy
Go Fish Digital discovered they could directly influence ChatGPT Search results through strategic content optimization. When analyzing how ChatGPT Search discussed their company for "digital PR agency" queries, they found a significant gap compared to competitors.
Key Findings:
ChatGPT Search was already citing one of their listicle articles frequently
Competitors were getting "Notable Clients" sections in AI responses
Go Fish Digital's listings lacked trust signals that competitors had
Implementation Details
Step 1: Content Identification
Located the existing article that ChatGPT Search was already using as a citation source
Confirmed this content was being scraped regularly by the AI system
Step 2: Strategic Content Addition
Added a "Notable Clients" section to the existing article
Used structured bullet lists to create key-value pairs that search engines and LLMs could easily parse
Formatted the information as data sources rather than narrative text
Step 3: Monitoring and Validation
Waited one week for the changes to be reflected in ChatGPT Search
Tested multiple variations of the "digital PR agency" query
Measurable Results
After 1 Week:
ChatGPT Search began pulling "Notable Clients" information when listing Go Fish Digital
Enhanced listing now includes trust signals previously missing
Improved competitive positioning in AI-generated responses
Business Impact:
Increased incoming business referencing ChatGPT as the discovery source
Enhanced brand credibility in AI search results
Established proof-of-concept for AI search optimization
Key Takeaway: "When ChatGPT uses the Search function, right now you CAN have the ability to influence the output. The first thing you need is an article/piece of content that can feed into the citations".
Derivate X: Achieving #1 ChatGPT Ranking for "Best Martech SEO Agency"
Company: Derivate X (SaaS SEO Agency)
Website: derivatex.agency
Challenge: Complete invisibility in AI search results for target keywords
Timeline: 1.5 months to achieve #1 ranking
The Comprehensive Strategy
Derivate X implemented a three-pronged approach focusing on topical authority, Bing optimization, and strategic mentions.
Implementation Details
Phase 1: Topical Authority Building
Created a comprehensive content web around martech SEO topics:
"SEO vs. Paid Ads: What's Better for MarTech SaaS Companies?"
"5 Common SEO Mistakes Martech Companies Make"
"Best Martech SEO Agencies in 2025: Top 10 Companies to Scale Your Business"
"Technical SEO Checklist for Martech Websites: What You Need to Know"
"What Role Can a CRM Play in an Effective Martech Stack?"
All content was strategically linked back to their main Martech SEO hub page.
Phase 2: Bing Optimization Focus
Recognizing that "optimizing for Bing is the closest thing to optimizing for ChatGPT", they implemented:
Tightened on-page SEO with perfect headings and internal linking
Comprehensive schema markup implementation
Site speed optimization
Keyword precision targeting
Phase 3: Authority Through Strategic Mentions
Instead of traditional backlink building, they focused on:
Engaging in relevant Reddit discussions
Contributing valuable answers on Quora
Publishing thought leadership on Medium
Building organic mentions in industry conversations
Measurable Results
Ranking Achievement:
Moved from invisible to #1 ranking on ChatGPT for "Best Martech SEO Agency"
Achieved top position in 1.5 months
Traffic and Engagement:
30-35 SaaS founders visiting their Martech SEO page monthly
3-6 calls booked per month directly from this ranking
2 new clients onboarded (Gumlet + one under NDA)
Validation Methods:
Reddit community confirmation of ChatGPT rankings
Direct client feedback about discovery through ChatGPT searches
AI system suggestions in related conversations
Conclusion
Transforming your content library for AEO and GEO is a strategic investment in future-proofing your search presence.
It aligns your content with the reality of how users seek information today – expecting immediate, AI-ready answers and AI-curated advice.
This playbook outlined how to audit your existing content, prioritize efficiently, rewrite for maximum answer impact, add the technical markup that signals relevance, and continuously iterate based on results.
Loved this post?
If you’re not a subscriber, here’s what you missed earlier:
The Lenny Rachitsky Playbook : Prompts, Growth Frameworks, and Strategies - Part 1 of 2
5M$ ARR with 6 people team - How Adam Robinson bootstrapped RB2B - Part 1 of 2
Founder-led Sales : Strategy and Guide (with AI Prompts) - Part 1 of 3
AI (ChatGPT, Gemini, Claude) Prompts for CMOs, Marketers and Growth Builders - Part 1 of 3
Subscribe to get access to the latest marketing, strategy and go-to-market techniques . Follow me on Linkedin and Twitter.
