Kasim Aslam's research team spent a year tracking how AI answer engines select their sources. What percentage of websites does his data say are using schema markup correctly, and how much more likely are those sites to get cited by AI?

Answer at bottom of email

Kasim Aslam used to run one of the world's top PPC agencies, Solutions 8. He managed over $100 million in ad spend. He's the kind of guy who should care deeply about Google rankings.

He spent last year building a six-person research team that does nothing but track how LLMs (large language models like ChatGPT and Perplexity) choose their sources.

What they found should rattle every marketer.

Voice schema, the structured data that helps voice assistants answer questions from your content, is used by a mere 0.3% of all websites.

Meanwhile, 55% of people under 25 use voice search daily.

The math isn't complicated. 90% of websites are effectively invisible to the emerging AI search ecosystem. Not because their content is bad. Because AI crawlers can't interpret it cleanly.

Kasim went further. He described a Google move that most marketers completely missed: Google quietly restricted the sources it feeds into AI training pipelines from "hundreds" to "ten." That single algorithmic shift cut off Reddit communities, YouTube discussions, and long-tail content from AI training. It effectively handed AI visibility to the same big publishers who already dominate traditional search.

His warning: "The K-shaped economy applies to AEO. Twenty percent of businesses will accelerate wildly. Eighty percent will fall to zero."

Then he said something even more important. Small businesses actually have an advantage right now. AEO rewards niche granularity and depth of expertise, not scale. The entrepreneur who goes deep on one specific topic, builds real schema infrastructure, and earns citations across platforms they already participate in will outrank the corporate brand that's still arguing about budget approval for a schema implementation.

The window is open. It won't stay open forever.

The brands that win AI search aren't the ones churning the most content. They're the ones who built the deepest expertise signals, implemented clean schema infrastructure, and show up authentically on the platforms AI systems already trust.

Marketer's Notepad:

It’s one of the oldest copywriting frameworks ever devised. Still one of the most effective for email, landing pages, ads, and content.

A. Attention. The hook. Your first line, headline, or opening image must stop the scroll. "Are you paying for the wrong keywords?" is Attention. "Welcome to our website" is not.

I. Interest. Give them a reason to keep reading. Reference their world, their specific problem, their industry. Generic interests nobody. Specific interests everybody in your target market.

D. Desire. Transform features into emotions. "48-hour delivery" is a feature. "Get it before your boss asks why it's not done yet" is desire. Desire is always about the outcome, not the mechanism.

A. Action. One clear, specific call to action. Not "visit our website." Not "learn more." "Click here to see how Kasim's schema audit template finds your AEO gaps in 20 minutes" is an action with specificity and benefit baked in.

Where it fails: Most marketers nail Attention and skip straight to Action. The Interest and Desire steps are where you earn the right to ask for the click. Don't rush them.

Segmented email campaigns generate 14.31% higher open rates and over 100% higher click-through rates compared to batch-and-blast sends. Yet most brands still treat their list as a single audience.

The most important segmentation shift right now is moving from demographics to behavior.

Start by tracking RFM scores (Recency, Frequency, Monetary value) and create distinct strategies for each tier.

High-value recent buyers get early access and loyalty rewards. Lapsed high-value customers get win-back sequences. New subscribers get onboarding flows.

For B2B, combine job title, company size, and product usage to form hyper-targeted segments.

Start with two or three behavioral segments and expand from there. The revenue lift is immediate.

For help with email marketing for your brand, contact dragon.fish

CheckThat.ai is a tool that shows you exactly how AI engines perceive your brand right now.

Paste in your brand name, products, or key topics and see how ChatGPT, Perplexity, and other AI answer engines describe you, rank you, and compare you to competitors. For AEO work, this is your baseline measurement tool.

You can't improve what you haven't measured. Run your brand through CheckThat before and after any schema or content changes to track your AEO progress. Free tier available.

The Brand: Liquid Death, founded 2018 by Mike Cessario.

The Strategy: Before a single can was manufactured, Mike ran Facebook ads for Liquid Death: just the branding concept, edgy design, aggressive tagline ("Murder Your Thirst"), to validate demand. He had zero product available.

100% brand. The ads proved that people would respond to water positioned as a rebellious, anti-plastic, punk-rock category disruptor. Then he made the product.

The Result: Liquid Death raised $1.6M in seed funding partly off the back of the pre-launch social proof. By 2024, the brand was valued at $1.4 billion. They became the fastest-growing non-alcoholic beverage brand in history.

The Lesson: Test the brand before the product. If you can't generate desire with pure positioning and no product, you don't have a product problem. You have a story problem. Fix the story first.

The cigar band, that decorative paper ring you slide off before lighting up, was reportedly invented by a European aristocrat who needed to keep his white gloves from getting stained with tobacco oils.

He wrapped a small strip of paper around the cigar's body as a barrier. What started as a laundry solution became one of the most recognized luxury branding elements in the world.

Today, cigar bands are collector's items. Some vintage bands from rare Cuban cigars sell for hundreds of dollars each. A practical problem, solved with paper, became a centuries-long brand icon.

The next Collective Minds Society Cigar & Whisky trip is Feb 18-22, 2027

Most SEO content is written for Google bots from 2015. AI answer engines read differently, select differently, and cite differently.

Here's how to close the gap quickly.

Re-run your AI citation check in 30 days to measure improvement. Sample Prompt:

"What questions would someone ask an AI assistant before hiring a [your service type] business? Give me 10 specific questions I should have FAQs for on my website."

The brands that make their content easy for AI to extract and cite will own the next wave of search visibility.

FAQ schema is one of the highest-return schema types for AEO because it maps directly to how AI engines operate: find a question, find a reliable answer, cite the source.

To get the most AEO value, your FAQ schema should follow the "answer first" principle. Provide a complete, concise answer of 40-80 words directly in the markup, not just a teaser.

Structure FAQ headings as real questions people actually ask. Use Google's "People Also Ask" and AnswerThePublic to find genuine query phrasing. After the short direct answer, provide a longer explanatory body on the page.

Keep questions specific, not broad. "What is the best time to send email for e-commerce?" will earn more AI citations than "How does email work?" Verify your implementation with Google's Rich Results

Test.

For help with this, contact dragon.fish

Only 12.4% of websites use schema markup correctly, and those sites earn 2.8 times more AI citations than those without it.

Kasim Aslam's six-person research team spent a year tracking how AI answer engines choose their sources and this was among their most striking findings.

Watch the full conversation with Kasim