For many marketing professionals, it might seem like online discovery shifted overnight. Gone are the days of scanning the search results, with users instead increasingly choosing to accept the search summary as their answer. Pew Research’s analysis of Google searches found AI summaries appeared in roughly one-in-five searches overall, and nearly three times more often for longer, question-like queries. With an AI summary on the page, users click a traditional result in fewer than 10% of visits, and click a cited source link just 1% of the time.
“Many marketing teams I’ve spoken to have felt the traffic changes,” says Wendi Lu, Chief Marketing Officer at Martinsen Global. “What’s easier to miss are the changes in credibility. With the first impression being mediated by an algorithm that lives outside your control, you’re at the mercy of the AI summary, and whatever verdict it assembles from the sources you didn’t vet.”
Lu, who advises leading international enterprise companies on digital brand strategy, joined us to unpack what this means for marketing and brand leadership. She also serves as a peer reviewer for IRJEMS and SARC Publisher journals on tech-forward, cross-cultural branding research. She argues that the emerging field of Generative Engine Optimization, or GEO, calls for a new branding strategy which blends classic marketing fundamentals with the realities of answer-first interfaces.
“Answer-first” feels more like a UX trend. Why should marketing leaders treat it as a brand issue?
AI is now doing part of the job your brand used to do. In a browsing model, you earned attention by pulling people onto your site, into your content, into your proof points; and that’s why ads grew into such a massive industry.
In the answer-first model, the system will synthesize the proof from sources you can’t fully curate, and may deliver a conclusion that’s incomplete or just not favorable.
Credibility can get established before a customer interacts with your brand directly. If the answer layer confidently names a dozen vendors with a few sources, many users treat that as enough to make a shortlist decision. People often end the session right there.
What might brands actually lose when they’re excluded from AI-generated answers?
The first risk is not being considered at all. If you’re not named in the top options, whether it’s the best nearby coffee shops or the best tools in a category, you’re stuck outside the decision set.
Over time, the brand that appears most often becomes the defaults, and the attention snowballs, simply because repetition is the fastest path to familiarity and trust. If inbound quality is softening or price pressure is rising, it can be a sign that buyers are coming to you with someone else framed as the standard. Marketing then has to spend to reintroduce a brand that should have been present at the first impression.
In your Forbes’ Editor’s Choice article, you wrote about how being considered is more than a matter of accuracy. What’s the biggest risk?
The term is context collapse. Conventionally, it refers to audiences being flattened into a single context, but that’s what these generative systems build from. The LLM behind the scenes will compress what’s most legible and high-signal about your category into a single narrative, regardless of where it comes from on the internet.
Tradeoffs and edge cases can get glossed over, so what’s best for one type of customer can be misrepresented as best for everyone. On top of that, the compression tends to favor brands that already dominate reference sources or show up frequently across widely cited sites, because they become the safest ‘consensus’ answer. So, the danger isn’t just an incorrect statement, but an incomplete narrative that nudges your customers away.
Where do citations fit into brand credibility? Why are they suddenly so important?
Some engines make this explicit. ChatGPT’s search experience can also provide answers with links to referenced web sources. Perplexity builds answers with numbered citations linking to sources, inviting users to do some homework. We’re generally moving towards better, more explainable AI, which is incredibly important.
But citations also function like authority signals. If an engine regularly cites a competitor’s materials, they can accrue credibility at the first impression. Being absent from citations means you might appear as non-authoritative inside the answer layers, even if you’re well-known elsewhere.
How about SEO? Does GEO replace it?
Search engine optimization is still an important pillar of your brand strategy. Today, at least, it feeds the retrieval layer behind those summaries. One study found that more than 75% of Google’s AI Overview-cited pages pull from the top organic results, so strong SEO still increases your chances of being cited.
The nuance is asking how it summarizes and credits your brand after retrieval. GEO is what governs this layer, and getting named and framed correctly in the answer people accept as their ‘completed search’ is just as important, if not more valuable.
What should marketing leaders do today to start responding to these changes?
Ask around, run the queries yourself. We refer to it as an answer layer audit, where you identify your customer’s top twenty or so prompts and document how your brand is represented.
Then start developing answer-ready assets informed by what you see. Build a one sentence description that establishes your brand’s canon, three to five proof points, then cap it off with a short paragraph you’d be comfortable seeing summarized. Inject those assets wherever you can, across your site and newsroom presence.
Finally, for brands in more technical spaces: strengthen your reference footprint. Publishing or earning materials that are easy to cite, like explainers and specifications, will give search engines scaffolding to pull from.
Lastly, do you believe GEO is here to stay?
Absolutely. When a platform as large and user-driven as Shopify starts publishing a GEO playbook, it’s a signal that answer-first discovery is already affecting how brands get found. The destination page is increasingly secondary to the prime real estate of the answer layer, and major platforms are investing in that interface as the default discovery experience.
