AI & Automation
Personas
SaaS & Startup
Time to ROI
Medium-term (3-6 months)
Last month, I was working on a complete SEO overhaul for a B2B startup client when something unexpected happened. While researching keywords using my usual tools like Ahrefs and SEMrush, I discovered their content was already appearing in AI-generated responses on Perplexity AI – despite being in a niche where LLM usage isn't common.
This discovery led me down the rabbit hole of what I now call GEO (Generative Engine Optimization) – and it completely changed how I think about content visibility in 2025.
Here's the uncomfortable truth: while everyone's obsessing over Google rankings, AI search engines like Perplexity are quietly becoming the new front door to information. Even in traditional industries, I'm tracking dozens of LLM mentions monthly for clients who never optimized for it.
In this playbook, you'll discover:
Why traditional SEO tactics fail spectacularly on Perplexity AI
The chunk-level content strategy that actually gets you mentioned
How to structure content so AI systems can easily extract and cite it
The 5 optimization layers that make your content citation-worthy
Real metrics from implementing GEO alongside traditional SEO
Let's dive into what everyone's getting wrong about ranking on AI search platforms.
Industry Reality
What every marketer thinks they know about AI search
Walk into any marketing conference today, and you'll hear the same tired advice about "AI-proofing" your SEO strategy. The consensus seems to be: keep doing traditional SEO, but maybe add some FAQ sections.
Here's what the industry typically recommends for AI search optimization:
Focus on featured snippets – because if Google shows it, AI will use it
Add structured data – schema markup will help AI understand your content
Create comprehensive FAQ pages – AI loves question-answer formats
Optimize for voice search – conversational queries are the future
Write for E-A-T – expertise, authority, and trust matter to AI
This conventional wisdom exists because most marketers are applying old-school SEO thinking to new-school AI systems. They're treating Perplexity like it's just "Google with a chatbot interface."
But here's where this approach falls short: AI systems don't consume content the same way search engines do. While Google crawls pages and ranks them based on hundreds of factors, AI models break content into passages, analyze context, and synthesize answers from multiple sources simultaneously.
The result? Companies spending months optimizing for featured snippets while their competitors get mentioned in AI responses through completely different content strategies. Traditional SEO metrics become meaningless when your content gets cited without anyone ever visiting your site.
What I discovered working with that B2B startup was something entirely different – a content approach that aligns with how AI actually processes and retrieves information.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
The client was a B2B startup in the project management space – not exactly what you'd call an "AI-forward" industry. Most of their prospects were still using spreadsheets and email chains to manage projects. Yet somehow, their content was showing up in Perplexity responses.
I'll be honest – this wasn't intentional. While doing keyword research using Perplexity instead of expensive SEO tools, I started noticing their company mentioned in responses to queries like "best project management workflow for remote teams." This was happening despite zero AI optimization efforts.
My curiosity was piqued. How was their content getting pulled into AI responses when they hadn't optimized for it? More importantly, could we replicate and scale this accidental success?
I started tracking LLM mentions across different AI platforms – Perplexity, Claude, ChatGPT, and others. Even in this traditional niche, we were seeing couple dozen mentions monthly. These weren't coming from aggressive GEO tactics but from solid content fundamentals that naturally aligned with how AI systems process information.
This discovery led me to conversations with teams at AI-first startups, and I realized everyone was figuring this out in real-time. There was no definitive playbook yet. What we did know was that the foundation hadn't changed – LLM robots still need to crawl and index content, and quality relevance remains the cornerstone.
But there was a new layer: chunk-level thinking. Unlike traditional search engines that evaluate entire pages, AI models break content into passages and synthesize answers from multiple sources. This meant restructuring content so each section could stand alone as a valuable snippet.
The breakthrough came when I stopped trying to optimize "pages" and started optimizing "information chunks" instead.
Here's my playbook
What I ended up doing and the results.
Instead of abandoning traditional SEO for shiny new tactics, I developed a layered approach that builds GEO on top of strong SEO fundamentals. Here's the exact framework I implemented:
Layer 1: Content Foundation (Traditional SEO First)
Before any AI optimization, we focused on creating genuinely useful content for humans. This remains priority #1. LLM robots still need to crawl and index your content, so traditional SEO best practices are your starting point. Quality, relevant content that serves user intent is non-negotiable.
Layer 2: Chunk-Level Optimization
This is where things get different. I restructured existing content so each section could stand alone as a valuable snippet. Instead of flowing narratives, we created modular content blocks. Each paragraph needed to be self-contained with enough context to be extracted and cited independently.
For example, instead of writing: "As mentioned above, this approach also helps with user retention..." we'd write: "The chunk-level content strategy improves user retention by 40% because each section provides immediate value without requiring context from other sections."
Layer 3: Answer Synthesis Readiness
We structured content with logical hierarchy that makes it easy for AI to extract and synthesize. This meant using clear headings, numbered lists, and step-by-step processes. But more importantly, we ensured each section could logically connect to related topics.
Layer 4: Citation-Worthiness Through Factual Accuracy
AI systems prefer to cite authoritative, factual content. We focused on including specific metrics, case studies, and data points that AI could confidently reference. Vague statements like "many companies see improvements" became "73% of B2B startups implementing this framework see 40% improvement in user retention within 90 days."
Layer 5: Multi-Modal Content Integration
We enhanced text content with charts, tables, and visual elements that AI could reference alongside written content. This created richer context for AI systems to understand and cite our information.
The key insight? Don't abandon what works. Build your GEO strategy on top of strong SEO fundamentals, not instead of them. The landscape is evolving too quickly to bet everything on optimization tactics that might be obsolete in six months.
What moved the needle wasn't aggressive GEO tactics – it was focusing on traditional SEO first, then adding the AI layer on top.
Chunk Thinking
Break content into self-contained, citable passages that work independently
Answer Structure
Use clear hierarchies and logical flows that AI can easily follow and extract
Citation Data
Include specific metrics and factual statements that AI systems can confidently reference
Multi-Modal
Enhance text with charts and visuals that provide richer context for AI understanding
The results from this layered approach were encouraging, though not revolutionary. Our tracked LLM mentions increased from couple dozen to consistent monthly visibility across multiple AI platforms. More importantly, the content that got mentioned was now intentional rather than accidental.
What surprised me was the timeline. Unlike traditional SEO that can take 3-6 months to show results, AI mentions started appearing within 4-6 weeks of implementing the chunk-level optimization. This faster feedback loop made it easier to iterate and improve.
The most valuable outcome wasn't the mentions themselves – it was learning how to think about content differently. Each piece of content became a library of citable information rather than a single page competing for rankings.
We also discovered that content optimized for AI performed better in traditional search. The clear structure, factual accuracy, and comprehensive coverage that AI systems favor also align with Google's content quality guidelines. It was a win-win optimization approach.
However, the landscape is evolving rapidly. What works today might be obsolete in six months, which is why building on solid SEO fundamentals remains crucial.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
Here are the key lessons from implementing GEO alongside traditional SEO:
Foundation First – AI optimization without solid SEO fundamentals is building on sand
Chunk-Level Thinking – Structure content so each section provides value independently
Synthesis Readiness – Make it easy for AI to extract and combine your information
Citation-Worthy Accuracy – Include specific data points AI can confidently reference
Multi-Modal Integration – Enhance text with visual elements for richer context
Rapid Iteration – AI mentions appear faster than traditional rankings, enabling quicker optimization cycles
Future-Proofing – Build flexible content strategies that work across multiple AI platforms
The biggest mistake I see companies making is treating this like a completely separate strategy. GEO works best as an enhancement to proven SEO practices, not a replacement.
What I'd do differently: Start tracking AI mentions from day one, even before optimization. Understanding your baseline helps measure improvement and identify which content naturally resonates with AI systems.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS startups implementing this playbook:
Focus on feature documentation with clear, self-contained explanations
Create comparison content with specific metrics and factual data points
Structure integration guides as modular, citation-ready information chunks
Track mentions across multiple AI platforms to understand reach
For your Ecommerce store
For ecommerce stores implementing this approach:
Optimize product descriptions with standalone, factual specifications
Create buying guides with clear, extractable decision criteria
Structure category content as comprehensive, citable information resources
Use customer data and reviews as authoritative, citation-worthy content