Growth & Strategy
Personas
SaaS & Startup
Time to ROI
Medium-term (3-6 months)
Three months ago, while working on an e-commerce SEO overhaul for a niche client, something unexpected happened. Our content started appearing in AI-generated responses—specifically in Perplexity AI—despite being in an industry where LLM usage isn't common.
This discovery led me down a rabbit hole that most marketers are completely ignoring. While everyone's obsessing over ChatGPT optimization and "prompt engineering," there's a massive opportunity hiding in plain sight: Generative Engine Optimization (GEO) for Perplexity AI.
Here's the thing most people don't realize: Perplexity isn't just another ChatGPT clone. It's fundamentally different in how it processes, cites, and presents information. And after months of experimenting across multiple client projects, I've cracked the code on getting consistent mentions.
In this playbook, you'll discover:
Why traditional SEO tactics fail in the AI-first search era
The chunk-level thinking approach that actually works for AI indexing
My 5-layer optimization system that generated dozens of Perplexity mentions
Content structure techniques that make AI systems love your content
Why most GEO advice is wrong and what actually moves the needle
If you're still thinking about SEO the old way, you're about to get left behind. This isn't about gaming the system—it's about adapting to how AI actually works.
Reality Check
What the AI optimization gurus aren't telling you
Walk into any marketing conference today, and you'll hear the same tired advice about "optimizing for AI." The SEO community has collectively decided that prompt engineering is the future, while content strategists are busy creating "AI-friendly" listicles.
Here's what the industry typically recommends for AI optimization:
Focus on featured snippets because "AI systems pull from them"
Use FAQ schemas extensively to appear in AI responses
Write in a conversational tone that mimics AI assistant responses
Target long-tail questions that people ask AI tools
Optimize for "People Also Ask" sections as AI training data
This conventional wisdom exists because most marketers are treating AI optimization like an extension of traditional SEO. They're applying old frameworks to new technology, assuming that what worked for Google will work for Perplexity, Claude, or ChatGPT.
But here's where this approach falls short: AI systems don't consume content the same way search engines do. They don't just look at title tags and meta descriptions. They break content into passages, analyze context, and synthesize answers from multiple sources simultaneously.
While everyone else is optimizing for yesterday's algorithms, there's a completely different game being played. And most businesses are losing because they don't even know the rules have changed.
Consider me as your business complice.
7 years of freelance experience working with SaaS and Ecommerce brands.
The discovery happened by accident. I was working on a complete SEO overhaul for an e-commerce client in a traditional, non-tech industry. The kind of business where you wouldn't expect AI usage to be common—think handmade goods and artisanal products.
About two months into the project, something interesting started showing up in our analytics. We were getting traffic from sources that didn't quite fit the usual patterns. The referring URLs were sparse, and the user behavior suggested people were arriving with very specific intent.
That's when I decided to test something. I started manually querying our client's industry topics in Perplexity AI, and there it was: our content was being cited in AI-generated responses. Not just once or twice, but consistently across different types of queries.
This was fascinating because we hadn't optimized for AI at all. We were just following solid content fundamentals—comprehensive coverage, factual accuracy, clear structure. But somehow, we were getting a couple dozen LLM mentions per month in a niche where AI usage "shouldn't" be happening.
Through conversations with teams at AI-first startups like Profound and Athena, I realized something important: everyone is still figuring this out. There's no definitive playbook yet. The industry "experts" are mostly theorizing based on traditional SEO knowledge.
This was my chance to experiment from the ground up, using real client projects as testing grounds. Instead of following the crowd's advice about "AI-friendly content," I decided to reverse-engineer what was actually working and build a systematic approach from there.
The challenge was clear: how do you optimize for AI systems that process information fundamentally differently than traditional search engines?
Here's my playbook
What I ended up doing and the results.
Instead of following conventional AI optimization advice, I took a completely different approach. I started by analyzing exactly how Perplexity AI processes and presents information, then built a systematic optimization framework around those insights.
Layer 1: Chunk-Level Content Architecture
The first breakthrough came from understanding that AI systems don't consume pages—they consume passages. I restructured content so each section could stand alone as a valuable snippet, following what I call "chunk-level thinking."
For my e-commerce client, instead of writing flowing articles, I created content where every paragraph answered a specific sub-question. Each section included relevant context, making it retrievable and useful even when pulled out of the full article.
Layer 2: Answer Synthesis Readiness
I discovered that Perplexity favors content that's structured for easy synthesis. This meant creating logical hierarchies, using clear subheadings, and presenting information in formats that AI can easily parse and combine with other sources.
We implemented numbered processes, bulleted key points, and comparison tables that made our content perfect for AI synthesis. The goal wasn't just to rank—it was to become the most citable source in our niche.
Layer 3: Citation-Worthiness Framework
Through experimentation across multiple client projects, I identified what makes content "citation-worthy" for AI systems. It's not about keyword density or traditional SEO signals. It's about factual accuracy, clear attribution, and comprehensive coverage that fills knowledge gaps.
We focused on creating authoritative content that other sources couldn't easily replicate. Original research, unique insights from industry experience, and comprehensive guides that became go-to resources in the niche.
Layer 4: Multi-Modal Integration
Unlike traditional SEO, AI systems can process and understand multiple content types simultaneously. I integrated charts, tables, and structured data that enhanced the content's value for both human readers and AI processing.
For the e-commerce client, this meant creating product comparison tables, process flowcharts, and data visualizations that made complex information digestible and highly citable.
Layer 5: Semantic Depth and Breadth
The final layer involved ensuring topical comprehensiveness. Instead of targeting single keywords, we covered entire topic clusters with semantic depth that positioned us as the definitive source on specific subjects.
This approach paid off when we started seeing mentions not just for our primary topics, but for related concepts and questions that we'd covered comprehensively in our content ecosystem.
Chunk Architecture
Structure every section to stand alone as a complete, valuable answer that AI can extract and cite independently.
Synthesis Ready
Format information in logical hierarchies that AI systems can easily parse, combine, and present to users.
Citation Quality
Focus on factual accuracy and unique insights that make your content worth citing over competitors.
Multi-Modal Value
Integrate charts, tables, and structured data that enhance content value for both humans and AI processing.
The results were more significant than I expected. Within three months of implementing the GEO optimization framework, we achieved dozens of Perplexity AI mentions per month across multiple client projects.
But the impact went beyond just AI citations. The approach created a compound effect:
Traditional SEO Performance: The content optimized for AI also performed exceptionally well in traditional search, with improved rankings and featured snippet captures.
Content Authority: By becoming the most citable source in our niche, we established thought leadership that extended beyond search visibility.
User Engagement: The chunk-level content structure improved on-page metrics, with users finding exactly what they needed quickly.
The timeline was surprisingly fast. Unlike traditional SEO that can take 6-12 months to show results, AI optimization started generating mentions within 4-6 weeks of implementation.
Most importantly, this wasn't a one-time win. The optimization framework created sustainable visibility as AI usage continues to grow across all industries and user types.
What I've learned and the mistakes I've made.
Sharing so you don't make them.
After optimizing for Perplexity AI across multiple projects, here are the key lessons that most marketers are missing:
Foundation Still Matters: Traditional SEO best practices aren't obsolete—they're your starting point. AI optimization builds on solid content fundamentals, not instead of them.
Quality Over Tactics: The biggest mistakes happen when people try to "hack" AI systems instead of creating genuinely valuable content that deserves to be cited.
Chunk-Level Thinking: Stop thinking in terms of full pages. AI systems consume and cite individual passages, so every section needs to be self-contained and valuable.
Context is King: AI systems reward comprehensive coverage over keyword targeting. Cover topic clusters with semantic depth rather than chasing individual search terms.
Multi-Modal Integration: Text-only optimization is incomplete. Charts, tables, and structured data significantly improve citation potential.
Speed Advantage: AI optimization shows results faster than traditional SEO, but requires consistent content quality and structure.
Platform Differences: Each AI system has unique preferences. What works for Perplexity might not work for ChatGPT or Claude. Test across platforms.
The landscape is evolving too quickly to bet everything on optimization tactics that might be obsolete in six months. Focus on building content that's valuable regardless of how AI systems change.
How you can adapt this to your Business
My playbook, condensed for your use case.
For your SaaS / Startup
For SaaS startups looking to get featured in Perplexity AI:
Focus on use case content that explains specific problems your product solves
Create integration guides that provide step-by-step technical value
Build comprehensive comparison content that positions you as an authority
Document your unique methodology in citable, chunk-ready formats
For your Ecommerce store
For e-commerce stores wanting Perplexity mentions:
Create buying guide content with comprehensive product comparisons
Build how-to content around product usage and care instructions
Develop industry knowledge hubs that establish topical authority
Structure product information in AI-friendly comparison tables and specs