Why Answer Engine Optimization Is Your New Search Strategy

In this week’s episode of Voices of Search, we spoke with Josh Blyskal, Head of AI Strategy and Research at Profound, about the seismic shift from SEO to GEO (Generative Engine Optimization) and why enterprise marketing teams need to act now before their competitors leave them behind. 

With ChatGPT visits growing 90% year-over-year to nearly six billion visits, the way people discover brands, products, and services has fundamentally changed. Josh revealed exclusive data insights from Profound’s analysis of over one billion citations and explained why the traditional blue link strategy is becoming obsolete in favor of contextual AI optimization.

Key Takeaways From this Episode:

• Blue links are no longer the primary unit of discovery, because users now make decisions inside AI answers where citations matter more than rankings.
• Answer engine optimization depends on context, not keyword density, and AI models reward content that clearly explains what a brand solves and why it’s relevant.
• Query fanout is the real search layer in AI systems, and brands need to optimize for the short, authoritative queries that models actually run—not the long prompts users type.
• Nearly half of AI interactions include unprompted product recommendations, which means brands can be discovered as solutions to problems before users show buying intent.
• Visibility in AI answers is built across every surface a brand touches, including PR, reviews, Reddit, Wikipedia, FAQs, and product data—not just the website.

The Fundamental Shift: From Hallways to Rooms

Search has evolved from being a hallway you pass through to reach other destinations into the room itself, where everything happens. As Josh explained, “Search has moved from being the hallway where you pass through it to get to these rooms to being the room itself. Search is where things happen. Search is the party.”

This transformation means marketers can no longer rely solely on traditional SEO tactics designed to drive traffic from Google’s blue links. Users are now solving problems, making decisions, and even completing purchases within conversational AI interfaces. The question isn’t whether this shift will happen—it’s whether your brand will be visible when it does.

Context Is the New King

If content was king in the SEO era, context has taken the throne in the age of answer engines. Profound’s research analyzing billions of citations reveals that success in AI discovery comes down to providing context that large language models can digest, understand, and use in their responses.

“Answer engine optimization is about providing context to models in a way that the models can digest, understand, and respond back to users with,” Josh noted. “We are in a high utility content world. The content that wins is not necessarily the same content that traditionally ranks in SEO.”

This shift requires marketers to think differently about how they structure and present information. It’s no longer enough to optimize for keywords and backlinks. Brands must contextualize their products, services, and expertise in ways that help AI engines understand not just what you offer, but why you’re the right solution for specific user problems.

Understanding the Information Portfolio

When we look at patterns in AI citations, sources like Wikipedia and Reddit aren’t just random domains getting lucky. They represent distinct classes of information that answer engines rely on to build comprehensive, trustworthy responses.

“Wikipedia, Reddit, these really represent classes of information,” Josh explained. “When ChatGPT increases Wikipedia citations, it’s not because it has some secret love affair with Wikipedia. It’s that they’re trying to index more heavily on pure factual, peer-reviewed encyclopedic knowledge. When they index heavier on Reddit, they’re testing how the influence of UGC impacts the quality, the factuality, and the perceived value of their responses.”

Currently, Wikipedia sits at around 7% of all ChatGPT citations, while Reddit hovers around 2-3%. These percentages matter less than what they represent: the need for both authoritative, factual information and authentic user-generated perspectives. Brands need to think about how they’re building their own diverse information portfolio across these different classes of content.

The Game-Changer: Query Fanout

Perhaps the most important concept Josh introduced is query fanout—the way AI engines simplify complex user prompts into concise, authoritative search queries. This is where the real optimization opportunity lies.

“You will put a 10 paragraph question into ChatGPT and ChatGPT is going to write three to five, three to five word statements,” Josh said, using the example of someone asking about running shoes with extensive personal details. ChatGPT distills that lengthy prompt into simple queries like “best men’s running shoes,” “best marathon running shoes,” and “running shoes store Brooklyn.”

This fanout process is visible in ChatGPT’s network logs and represents the actual searches the AI engine is running to build its response. “These are the authoritative actual queries,” Josh emphasized. “That’s where we need to intersect our content with. We need to literally aim our content to hit those fanouts because that’s the actual search.”

Understanding query fanout means you can:

• Identify which user details actually matter to the AI engine
• Discover the specific phrases you need to optimize for
• Avoid wasting resources on unnecessary content variations
• Test whether different personas actually trigger different fanouts (often they don’t)

The best part? You can access this data yourself. Josh outlined the process: go to ChatGPT’s network logs, copy the conversation ID from the URL, open the inspect element, filter by that ID in the network tab, and search for “queries” in the fetch requests. There you’ll see exactly which searches ChatGPT ran to build its response.

Why Prompt Volume Matters More Than You Think

While many marketers worry about the lack of data transparency from AI platforms, Profound has assembled what may be the industry’s most comprehensive dataset: over 150 million individual prompts collected monthly across multiple regions and models.

“The coolest thing that no one talks about—everyone is like, heuristically, culturally in the SEO space, ‘as we all know, nobody has the data,'” Josh said. “We have over 150 million individual prompts that we’re getting piped into our system every single month.”

This prompt volume data serves as the foundation for understanding what users are actually asking about your space, your competitors, and your industry. When combined with fanout analysis, it provides both the questions people ask and the simplified queries that drive citations.

The key is using prompt volume as a means to an end. “Prompt volumes are good only to get me to the fanouts,” Josh explained. “Everything else exists in service of the fanout.” This multi-gate system—understanding the original prompt, tracking how it evolves through the system, and identifying which sources are cited—provides marketers with a comprehensive view of the AI discovery journey.

The 47% That Changes Everything

One of the most surprising findings from Profound’s research: 47% of conversational AI interactions feature unprompted product recommendations. Users don’t always start by saying, “I want to buy X.” They start with problems.

“People actually come in and say, ‘I need a new desk setup. My back hurts.’ And the response is ‘Oh, have you considered Herman Miller?'” Josh illustrated. “It doesn’t mean that people aren’t shopping. It means that people’s shopping journeys start with what’s really core to them. This is my need. This is what needs to be solved.”

This represents a massive opportunity for brands. You’re no longer waiting for users to search your brand name or product category. You can be recommended as the solution to problems you might never have associated with traditional search keywords. The key is contextualizing your brand, product, or service in that gap between problem and solution.

Multi-Surface Consensus Building

One of the hardest truths for marketing teams to accept is that AI optimization can no longer be siloed within the search team. Every surface your brand inhabits online now contributes to your visibility in AI responses.

“Every surface that your brand inhabits now online now trickles down to the end goal of visibility, whether or not you want it to,” Josh said bluntly. “You can either create and have that knowledge innately that everything we produce is now going to be factored into this all-knowing search system, or I’m just going to keep doing the thing I was doing and not really link in.”

This means:

• PR releases aren’t just announcements—they contextualize product accuracy in ChatGPT for months or years
• Social media activity on Reddit contributes to brand sentiment and authority
• Affiliate content needs narrative consistency with your brand messaging
• Wikipedia pages provide legitimacy signals
• User reviews across platforms feed into AI understanding

For VPs of growth and marketing leaders, this requires horizontal coordination across teams that traditionally operated independently. As Josh put it, you’re “building consensus through these multi-surface consensus-building activities.”

What Works Right Now (And Why It Won’t Last)

We’re still in what Josh calls “turn one” of AI optimization strategies. Listicles currently dominate AI citations because they provide exactly what answer engines want: comprehensive, structured information that’s easy to parse and recent.

“The reason that works so well is because these answer engines are lazy,” Josh explained. “They want to have someone do the legwork in the room of saying, ‘I don’t want to figure out every corporate card. I love this source. It was published three days ago. It’s got every single corporate card on there.”

But here’s the critical insight: as more brands adopt these strategies, their effectiveness will decline. Just like in traditional SEO, once everyone has optimized listicles, the next layer of differentiation will emerge. The key is to adopt what’s working now while preparing for the inevitable evolution.

Josh’s research also revealed surprising findings about backlinks in the AI era. Having two or three backlinks gets content into AI consideration sets, but the relationship plateaus quickly. More interesting: semantic mentions (when your brand is referenced without a hyperlink) carry roughly equal weight to traditional inline links in answer engine evaluation.

The Commerce Revolution Coming in 2026

While we’re in a testing phase for AI commerce, Josh predicts dramatic changes ahead. “Everything is going to come down in a year’s time when we’re thinking about commerce. We’re going to think the idea of web-based RAG for commerce is disgusting and gross.”

The future is backend integration. Instead of AI engines scraping websites to find product availability, they’ll connect directly to merchant systems through APIs and product feeds. This means the structured data work that SEOs have been doing for years suddenly becomes exponentially more valuable.

For the holiday shopping season currently underway, Josh’s team tracked dramatic differences in ChatGPT responses between the Friday before Black Friday and the Friday after. The AI engines are adapting in real-time to user behavior during high-intent shopping moments.

The practical advice? Start thinking about how to translate your offerings into product feed specifications. Even service-based businesses should consider structuring their offerings as JSON objects that AI engines can parse and recommend. As Josh put it, “You don’t have to upload your product feed just yet, but you do have to think eventually, when this starts to come through, how are you going to contextualize what that is going to look like for you?”

The FAQ Advantage Nobody’s Talking About

In analyzing 10,000 top-performing product detail pages versus 10,000 bottom-performing ones, Profound discovered that FAQ content was 848% more prevalent in the top-performing set. That’s an 8x to 9x higher chance that a top-performing PDP includes FAQ content.

This isn’t about randomly adding FAQs to check a box. It’s about providing the structured, question-answering format that AI engines naturally gravitate toward when building responses. The content must genuinely assist the AI in creating more effective answers.

“You want to be in lockstep with OpenAI,” Josh advised. “You want to think if I’m OpenAI, do I look at that piece of content and say, ‘Oh my goodness, I wish everyone created pieces of content like this. It makes ChatGPT’s answers so good.'”

Why SEOs Are Uniquely Positioned to Win

Despite all the change, Josh sees this as a massive opportunity for search professionals. “For SEOs, very clearly, this is a huge moment. CMOs, your C-suite, everybody is looking for someone to raise their hand and say I know about structured data and these scary systems that rely on so much structured data.”

The skills that made SEOs successful—understanding structured data, optimizing for intent, analyzing user behavior patterns—directly translate to answer engine optimization. The units are the same (URLs, schema, on-page content), they’re just being used in slightly different ways.

“We are doing the same exercises, we’re working out the same muscles, but we’re doing totally different exercises,” Josh explained. “We’re still working with URL slugs. That didn’t go anywhere. Schema is still here. You’re working on page content. All these things are still there.”

The opportunity is particularly strong for SEOs willing to bridge into product and engineering conversations about things like product feeds and backend integrations. Where else in an organization would you find someone who specializes in structured data and understands how to make information machine-readable?

The Attribution Problem (And Why Citations Matter)

One of the biggest challenges in AI optimization is attribution. Users often discover products in ChatGPT, then navigate to Google to complete the transaction. This makes traditional click tracking nearly impossible.

Josh predicts “there’s a multi-billion dollar startup sitting out there just to solve this attribution problem of people copy-pasting the name of the product they saw that was visible in ChatGPT into Google and then transacting there.”

In the meantime, citations provide the best proxy metric for impact. “Citations are the new clicks,” Josh said. “Citations are clicks, but from bots basically.” When a real person asks a question, and your content gets cited in the response, that’s a real moment where your brand provided value—regardless of whether you get a trackable click.

The key is calibration. Once you understand your baseline citation patterns for relevant queries, you can track improvements over time and correlate them with the content and optimization changes you’re making.

Getting Started: The Practical First Steps

For marketing teams feeling overwhelmed, Josh offered remarkably accessible advice for getting started:

Do manual testing: Search 10 queries in ChatGPT related to your space. Grab the sources cited. Look at which brands appear. Check the network logs for fanouts.

• Monitor and measure: Come back the next day and repeat. Track changes over time in a simple spreadsheet. Look for patterns in which content types and brands consistently appear.
• Test and iterate: Make changes to your pages based on what you learn. Within weeks, you may see citation improvements.
• Check your technical foundation: Use simple JavaScript tests to see if answer engines can render your content. Copilot and Gemini can now both render JavaScript, removing a major blocker for legacy sites.
• Start with one high-quality piece: Create one exceptionally well-contextualized piece of content—a comprehensive listicle, a detailed product landing page, or an FAQ-rich guide—and see how it performs.

“You don’t even need a tool suite at that level,” Josh emphasized. “You just need to understand basic marketing.” The tools become valuable when you want to scale, automate analysis, and manage optimization across hundreds or thousands of queries.

Why Now Is the Blue Ocean Moment

Despite all the anxiety in the marketing world about AI disruption, Josh sees this as a remarkably calm period for early adopters. “Right now, blue ocean, it’s really actually calmer than most people would have it believe. It’s a lot like bowling with the bumpers.”

We haven’t reached the point where everyone’s armed with the same data and executing the same strategies. That means the brands that move now have a significant first-mover advantage. They can establish authority in AI responses before the space becomes hyper-competitive.

“We are in turn one where people haven’t yet even consolidated onto the few things that are very clearly working in the space,” Josh explained. “Once we do that, things start to get more indirect. It starts to become about combining variables and testing multivariate statistical tests. It gets really in the weeds.”

The message is clear: the complexity will increase. The data will become more sophisticated. The optimization will require more advanced techniques. However, there’s low-hanging fruit for marketers willing to understand the basics of query fanout, contextual optimization, and structured data.

The Bottom Line

The shift from SEO to GEO isn’t coming—it’s here. With ChatGPT alone approaching six billion visits and AI discovery becoming the default way millions of users find information and make decisions, marketers can’t afford to wait.

The good news? The fundamental skills of search marketing still apply. The expertise around structured data, user intent, and content optimization translates directly. And the brands that invest now in understanding query fanout, building contextual authority, and creating high-utility content will establish positions that become harder to displace as the space matures.

As Josh put it: “Context is king and context is going to dictate success in this space in the next year for sure.” The question is whether your brand will provide that context or watch competitors fill the void.

The opportunity is massive. The time is now. For search professionals willing to take the lead in this transition, the moment represents one of the most significant career opportunities in the history of digital marketing.

Start with fanouts. Build your context. And remember: the answers are always in the data.

Voices of Search is a daily SEO and content marketing podcast hosted by Jordan Keone and Tyson Stockton.The show delivers actionable strategies and data-driven insights to help marketers navigate the ever-evolving world of search engine optimization and content marketing. New episodes air weekly, covering everything from technical SEO to AI discovery, featuring industry leaders and practitioners sharing real-world frameworks and proven tactics.

Subscribe to Voices of Search on Apple Podcasts, Spotify, or your favorite podcast platform. Follow Previsible on LinkedIn for updates and subscribe to the VOS YouTube channel for video episodes and clips. You can also visit the official VOS site to explore the full episode archive and submit your SEO questions for future episodes.