Why AEO Starts With Better SEO Fundamentals

In this week’s episode of Voices of Search, we spoke with Lucas Tieleman, CEO and co-founder of Optiversal. Lucas helps some of the world’s largest retailers optimize for AI-driven discovery by turning product data into machine-readable content that performs across both traditional search and emerging AI platforms.

Our conversation covered why AEO starts with better SEO fundamentals, how AI is changing the way people search rather than replacing search altogether, and why creating more content isn’t the same thing as creating more value.

Key Takeaways From This Episode:

• Search isn’t shrinking or splitting into winners and losers—the overall market is expanding as consumers ask longer, more natural questions across more platforms
• AI content tools remove the old trade-off between hero products and long-tail catalog, but scaling content isn’t the same as scaling value
• Product feeds now power Google’s product graph and emerging platforms like OpenAI’s commerce protocol, making feed strategy a genuine competitive differentiator
• Executive urgency around AI search is finally unlocking budget for fundamentals—like clean structured data—that were hard to prioritize a few years ago
• Structured data and clean entity relationships matter more today than they did five years ago, not less

Search Isn’t Shrinking, It’s Expanding

Lucas opened by dismantling a framing he hears constantly: that AI search is a zero-sum game, stealing share from traditional search until one platform wins and the other fades. 

The data doesn’t support it. Google search volume is up. AI Overviews have gone from appearing on roughly 6% of searches to 40% in the past six months. And new platforms are pulling in queries that wouldn’t have existed as searches before. 

“The pie is expanding,” Lucas said, “and there are many more interesting ways to search for things now.”

What’s actually changing, Lucas argued, is consumer behavior rather than platform dominance. People no longer compress their intent into a bare keyword and sort through filters afterward— they ask the fully specified question up front. 

“If you love coffee, you can actually ask very detailed questions about the espresso machine in the search,” he said, describing a shift from “espresso machine” to a query that already includes grind type, size, and brewing style—collapsing what used to be a multi-step research process into a single, upstream question.

AI Removes the Hero-Product Trade-Off, Removing Cost Constraints

This shift in consumer behavior changes what’s economically viable to write about. Retailers used to have to ration content quality: a handful of hero products got real investment, while the rest of the catalog got a sentence and a half. AI content tools remove that constraint by making high-quality content cheap enough to produce across an entire catalog, not just a chosen few SKUs.

This also opens up audiences that used to be too small to justify the spend. “You might have products in your catalog that have a unique audience, but that audience two years ago might not have been big enough to justify the marketing spend,” Lucas said. Now it is.

Customer-Backwards Beats Technology-Forward

Removing the cost constraint creates a new problem, however, as it’s now just as easy to generate 10,000 mediocre blog posts as it is to generate value. 

Lucas borrowed a framework often attributed to Jeff Bezos—building customer-backwards versus technology-forward—to draw the line. 

• Technology-forward means producing content because the tools now make it easy.
• Customer-backwards means starting with what the customer actually needs and using AI to meet that need at scale.

“Can we basically scale value versus just scaling content?” Lucas said. “That might work for a very short period, but almost always that doesn’t work over a long period of time.”

A Necklace That Changes With the Seasons

Lucas illustrated the customer-backwards approach with a jewelry retailer Optiversal works with. 

For instance, the same necklace sells year-round, but the reason someone buys it changes constantly: a Valentine’s Day gift in January, a Mother’s Day gift in May, a graduation gift by early summer. 

The brand used to run a single seasonal marketing campaign at a time. Now the product description, FAQs, and structured data update to match whichever moment is live, and the content shifts to match the customer’s actual reason for buying, down to the copy itself. 

“Mothers love it because of X, Y, and Z,” Lucas said, describing how the same product page reframes itself for graduation season a few months later—work that used to require a new campaign each time now happens as a continuous update to one asset.

Product Feeds Are Becoming Prime Real Estate

That same logic extends beyond the product page into the product feed itself, which Lucas described as increasingly central to how discovery actually works. Feeds power Google’s product graph and the systems behind its universal commerce protocol, which means normalizing a retailer’s internal taxonomy against Google’s own categories directly affects whether products get found at all.

Here’s what that loop looks like:

• Merchant feeds now surface FAQs directly in organic results, the same content shown under “what you want to know next” on a product listing
• AI Mode draws on that same feed data to answer product questions
• Meta and OpenAI are each building their own feed field requirements, turning feed strategy into a platform-by-platform discipline rather than a single one-size-fits-all export

“Really having a feed strategy beyond just making sure it gets to the place where it needs to get to is very important,” Lucas said.

Old Fundamentals, New Executive Budgets

One side effect of the AI search boom that Lucas finds genuinely useful: it’s suddenly much easier to get budget approved for unglamorous technical SEO work. A pitch to overhaul schema markup across an entire site used to be a hard sell. Once executives start asking why their brand isn’t the first result in ChatGPT, that same investment becomes urgent by comparison.

“Three years ago, if you were pitching to really update the schema at the highest quality level across all the pages in the company, that’s a tough roadmap decision,” Lucas said. “But now all of a sudden ChatGPT is here, AI Overviews are here, and with that come the budgets.”

Learn First, Then Test

Because content and experimentation are now cheap to produce at scale, Lucas argued for flipping the usual test-and-learn model on its head. Running a thousand experiments and hoping a few reach statistical significance isn’t a strategy just because the volume is now affordable. 

The better sequence starts with understanding the customer, then testing a specific hypothesis about what they need.

“Learn what the audience is saying, and then test if your validations are right,” Lucas said. The goal isn’t running fewer tests—it’s making sure each one is actually answering a real question instead of just seeing what sticks.

Retire “Sloptimization”

When asked what tactic marketers should abandon, Lucas didn’t hesitate: mass-producing keyword-stuffed articles by feeding research straight back into a content model. 

“If you are going to do a bunch of keyword research and turn it into 1,500-word posts based on the keywords you researched and just feed them back into the model, you are perpetuating the problem, and you’re not going to make more money,” he said.

His broader concern is that people still believe AI search behaves like the old keyword-ranking game, just with a new leaderboard. It doesn’t. Results are increasingly personalized, so two people searching the same phrase can see entirely different answers—which means the right question isn’t whether you rank first, but whether you have the right content for the person actually asking.

Talk About Photos, Not Gigabytes

When asked for the one piece of advice he’d give every e-commerce team, Lucas pointed away from technology entirely and back toward a habit he picked up working retail at Apple: sell the way you would if the customer were standing in front of you. Nobody walking into a store to buy a laptop for their kid’s photos cares about gigabytes. They care about how many pictures they can store.

That same instinct shows up across categories once retailers start writing from the customer’s actual problem instead of a spec sheet—pet food framed around a dog’s sensitive stomach, mascara framed around sensitive eyes. It’s the same shift Lucas described earlier with search queries themselves: customers aren’t leading with specs, they’re leading with the problem they’re trying to solve, and the content that wins is the content that meets them there.

“Customers search that way. They search from their problem,” Lucas said.

Why the AI Shopping Agent Isn’t Coming as Fast as You Think

When the conversation turned to what today’s AI search discourse will look like in hindsight, Lucas pointed to agentic commerce — AI agents that don’t just help you research a purchase, but complete it for you. He doesn’t think it’s a dead end, but he does think the timeline is off.

His evidence came from an ordinary Target run, standing in the detergent aisle. Every piece of infrastructure needed for an agent to reorder detergent automatically already exists: people are fiercely loyal to one brand, subscribe-and-save is everywhere, and it’s about as low-consideration a purchase as exists. And yet, the aisle is still full of people buying it themselves.

“If it doesn’t even work for detergent, how is it going to work for clothing or for large purchases?” he said. His read is that shopping carries an emotional component — the desire to be the one who chooses — that no amount of infrastructure resolves on its own, even at the highest end of retail. “I don’t think the AI is going to choose for you,” he said. “But maybe I’m getting old.”

The Old Playbook Isn’t Obsolete, It’s Overdue

Nothing in this conversation described a new game with new rules. Search is still won by being genuinely easy for a machine to read and genuinely useful to the human behind the query. Lucas just kept finding the same principle showing up in a new place, whether that was a merchant feed, a seasonal product page, or a piece of copy that talks about photos instead of gigabytes.

What’s changed isn’t the fundamentals. It’s that AI has made them cheap enough to apply everywhere at once, and urgent enough that leadership finally wants to pay for it.

Voices of Search is a daily SEO and content marketing podcast hosted by Jordan Koene 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.