Panic! at the GEO: How Do You Show Up in an AI Search Answer?

Search data firm Semrush predicts that AI engine queries will exceed traditional Google search queries some time in 2028. That has professional services CMOs and business development heads actively scrambling to ensure their firms earn a seat at the AI table. Google searches still currently dominate AI searches by an enormous margin, with search data firm Semrush reporting Google got about 14 billion searches per day in 2024, while ChatGPT only got about 37 million “search-like” daily queries. But the early warning signals of a sea change are there: Searches on Apple’s Safari browser fell for the first time ever. Survey results show user trust of Google search is dropping and being replaced by AI and social media. Young people are using AI at rates more than double that of older demographics, pointing to accelerating trouble ahead for traditional search engine marketing. The rapid improvement in the performance of AI platforms like ChatGPT, Claude, and related platforms like Perplexity are wreaking havoc with the way buyers find resources. Search behavior and AI engine query behavior differ—AI generally gets longer, more detailed queries than a Google search box and responds with fewer, more specialized answers or recommendations than Google’s multiple pages of search results. Showing up as one of those answers on the AI platform, or having your firm’s work cited and used as the basis for an answer, becomes key. The rising term for the art of improving the visibility and performance of a company on AI platforms is Generative Engine Optimization, or GEO. There’s enough “black box” functionality in AI large language models that exact prescriptions for GEO success remain elusive. But a good place to start is by understanding how LLMs determine what content should be prioritized in its answers. LLM authority frameworks are the mechanisms that help an AI model decide who to cite and trust. Each AI model will differ in the specifics of its authority framework, but here are the basic building blocks: Observation and statistical inference. Data and information that gets cited in multiple other places on the web implies that it is authoritative. A publication whose work is linked to frequently begins to accrue a higher authority rank, and content that appears in that publication is then treated accordingly. Factual alignment. Clusters of facts that continue to line up in the same way in multiple sources reflect well on the facts’ veracity, as well as on the publisher of those facts. Semantic clues. LLMs draw inferences from the language around particular facts and data, which means that a kernel of information that appears in a peer-reviewed academic article with accurate citations and more formal language will carry more weight than a similar bit that appears in an informal post on a public blog or social media site. Structural focus and specificity. LLMs can more easily show confidence in a source that uses short declarative sentences to express a focused set of ideas. The alternative—either more literary sentence structures with parentheticals and nested ideas, or the more colorful vernacular and opinion-driven sentences that might turn up in a comment section—are harder to machine-parse with high confidence levels. Supervised training. LLMs receive some human feedback later in the process, often in the form of ranked answers (“Which response below do you prefer?”) or direct human feedback to the model on what sources or publications are more authoritative. Some preexisting bias can creep in through such doors. But these are just the general outlines of LLM authority frameworks. AI platforms, in effect, tell no tales and don’t give up their black-box secrets. Sometimes that’s for proprietary reasons, but for the most part, it’s because it’s just very hard to trace how a model learns and “thinks.” Still, with the given authority framework components, the GEO dos and don’ts for marketers will be driven by maximizing a firm’s content and web presence in ways that the LLMs find reassuringly authoritative. Here are some touchstones in getting started. Publish thought leadership in high-authority locations This tactic is the most obvious and necessary element for GEO (including if you ask the models themselves what a marketer should be doing to drive AI presence). This is a difficult task though. Earning a spot in The Wall Street Journal isn’t an everyday occurrence, no matter how prestigious the firm. Quality, newsworthy content is hard to conceive, hard to execute, and very hard to publish. But it’s the most powerful GEO medicine, so consider the regular creation of incisive content for outside publication. Focus on trade journals, first- and second-tier business and financial publications, and metropolitan newspapers, which retain high authority even in their declining years. Create short, single-subject content It appears that making facts obviously discoverable and indexable is preferable to creating an all-encompassing guide to an infinitely broad topic. Your firm needs content that goes deep and stays narrow. Reinforce it through obvious titles, well-demarcated content sections, tight expert commentary, and well-cited, on-target data that backs up assertions. Partner with authoritative researchers and data sources Reflected light is better than no light at all. Use outside resources to help create genuinely original acts of expert journalism. The best way to show up on a focused topic is to create something containing genuinely new knowledge or insights. If you haven’t already, a good starting place is to perform a simple GEO self-audit: Search your firm, its partners, and its focus areas in various LLM tools to assess what is and isn’t being picked up about you, and how different LLM services are presenting (or misrepresenting) your firm. That gives you a good starting place to determine what kind of authoritative content your team can create or direct in order to fill gaps and reinforce positives. Moving your firm from a world that Googles to a world that collaborates with ChatGPT or Claude is a huge change. Your GEO success will depend on moving your content and ideas from your own website to more high-authority venues and making sure it’s accessible and comprehensive. Not an easy task, but getting a head start in these early days of AI adoption can help.

Voice

Panic! at the GEO: How Do You Show Up in an AI Search Answer?

Soaring generative engine queries have CMOs scrambling to get their brands cited

UNLOCK FULL ACCESS

Dave Maney

Performance Marketing

Trust and Incrementality Help Brands Scale Bigger Ideas

Entries Open for ADWEEK Tech Challengers to Watch 2027

How Marketers Can Deliver Outcomes Across Fragmented Journeys

A More Intelligent Advertising Ecosystem Is on the Horizon

When Past Purchases Aren’t Enough Anymore

Healthcare Marketers Rethink Infrastructure in Search of Real-Time Results

ADWEEK Innovator 50 Entries Now Open for 2026

Permutive Recruits Former IPG Mediabrands CEO to Board in Buy-Side Push

How Marketers Can Get Clean Incrementality Measurements in a Messy World