What AI Knows–and How Brand Marketers Can Teach It

Published on December 24, 2025

Learn the SEO signals GenAI uses for authority and trust.

The old model of SEO in PR was a technical game of keywords and backlinks. The new model of Generative Engine Optimization (GEO) is driven by artificial intelligence capabilities and Google’s E-E-A-T grading system for Experience, Expertise, Authority and Trust. Brand and content marketers who understand GEO can provide the clear, expert point of view that AI-powered search favors. 

Google’s AI Overviews in search engine page results built on Google past practices of building “knowledge graphs” of word relationships. AI opens up these indexes beyond keywords. For more complex queries, the Gemini-powered AI Mode puts research on automatic pilot, making a bigger and more ominous leap from AI-assisted search to search-assisted AI.

AI Dives Down the Content Marketing Funnel

Both AI-curated search and search-enhanced AI are growing sources of website traffic. ChatGPT and Perplexity web referrals now rival the volume of traffic from search engines like Microsoft Bing, Duck Duck Go and Yahoo. Brand marketers can track this growth, in the acquisition tab of Google Analytics, as the number and proportion of AI referrers over time. Choose a time frame and comparison period, then sort user acquisition metrics by first-user source. 

Google is making its biggest AI impact in three products:

• The AI Overview feature in Google search replaces featured snippets on search engine results pages. Its AI-generated summaries appear above the list of search links, with callouts to encourage users to make immediate follow-up queries.
• AI Mode largely abandons the standard SERP format with a chatbot interface that accepts voice and image inputs as well as text. AI composes longer query responses, with links presented held to a sidecar display. 
• Gemini, Google’s AI chatbot, produces long-form briefings similar to whitepapers, showing its sources as embedded links.

The growth of AI in search likely is being underestimated, because AI bots do not transmit referral information consistently. When analytics programs cannot determine AI as a traffic source, content marketers will see an increase in “direct” traffic instead. 

The rules of search optimization change when AI does all the searching. Keyword density isn’t dead, but its pulse is weak. Instead, GEO helps GenAI make semantic distinctions among topical keywords. The breakouts in Google Trends suggest how AI builds a knowledge graph of related concepts, filtering them by source, activity and region. This “Who, What and Where” analysis maps content to a query’s intent.

Backlinks aren’t exactly gone either, but the links themselves are optional. AI models are better at recording mentions across the internet–from name-dropping on message boards to citations in academic journals–and inferring the strength of each testimonial. GEO backs up copywriting claims with third-party endorsements and hard numbers on ROI metrics.

Open AI’s ChatGPT, Microsoft Copilot, Google AI Mode and other generative engines are thoroughly trained from web sources. But AI models do not “read” for inspiration; they parse for relationships between entities (for brand marketers, their clients), attributes (product features or personal expertise), and solutions (what they fix). Like product displays, sports scores and other past enhancements to search results, AI research draws from web source material but presents results directly rather than simply providing links.

By clicking through the web and briefing its users, Gen AI pulls them deep down the marketing funnel. They summarize a topic, then for purchase-intent queries, they go further to make product and service recommendations. As users shift from search queries to AI prompts, overall web views are likely to fall. Users most often will click directly from AI results to a recommended item description (often an affiliate link) or to a company contact or landing page. AI might also direct users to a LinkedIn profile or conference speaker’s bio to demonstrate a service provider’s authority.

Case Studies: Purpose Branding and Positioning

• Meeting the Moment for a Robust Public Health Regimen
• Educating and Engaging O’Hare Communities
• Social Media Amplifies Sweepstakes to Promote Debt-Free Education for All
• Healthcare Startup Embraces a Neglected Market

Winning the AI Recommendation Game

The users of generative engines increasingly ask their chatbots to solve problems and make purchase decisions. In essence, AI does the shopping for them, researching prices and features across web sources. A unique selling proposition must be not only intuitive to website visitors, but also clearly stated, structured and coded for computers.

AI builds the same kind of matrix a project manager might keep on a spreadsheet to compare vendors. The models infer product and service capabilities from multiple providers and give their users “warm” introductions of companies that align with not only the user’s stated needs but also their implied interests. 

Viewed this way, the opportunity to optimize for AI to present a differentiated solution with real-world results–a familiar branding assignment. Three steps can make this marketing message stick with both purchasers and their AI agents:

• Own the Definition. Generic descriptions (“healthcare consulting”) are only the first step in classifying an offering as clearly relevant. To recommend a specific solution, AI first needs a name for it–why not give it one? Brand your client’s products, events, services, topics and frameworks with useful shorthand terms (“fall risk assessments,” “business model redesign”). Provide a distinct or detailed description–a snippet in SEO-speak. It doesn’t have to answer a question–the game is clarity, not “Jeopardy!”
• Use Subject-Verb-Object Sentences. Generative engines seem to like their adjectives–every solution is “innovative,” “game-changing” or “transformational”–but they really are looking for the basic Who, What and Where of a proposition. “In-home assessments lowered ER visits 25%” is a statement that both potential clients and AI databases will read as powerful–even visionary. 
• Be the STAR of Your Story. Testimonials and case studies are their most effective in the storytelling format of Situation, Task, Action and Result. (Alternately, the SOAR format describes an Obstacle to overcome.) In AI comparisons, the results metric marks a trustworthy provider, and revenue analytics can be definitive. However, volume and other service metrics are solid demonstrations of scale and reliability. GenAI might quote satisfied clients but still rely on star ratings or other experiential benchmarks.

Case Studies: Purpose-Driven Brand Engagement

• The Placement With the Helpful Millennial Focus
• Staging a Welcoming Homecoming Journey
• Social Media Framework Promotes Gifts with Social Impact
• Three Public Transit Agencies. One New Payment Method. Millions of Customers Onboard.

Purpose Prompts: How AI Can Sharpen a Message

Content marketers should review their copy for concrete, AI-ready cues–and AI can help them do it. Here are prompts to test content for purpose-driven differentiation.

• Generic vs. Specific. “Act as a critical editor of the following text. Identify every adjective or phrase that could apply to any competitor. List specific phrases and suggest a concrete, data-backed alternative based on facts provided in the text, or flag the phrase as ‘Needs Evidence.'”
• Enter the Matrix. “Act as a knowledge graph builder for a search engine. Parse the text and extract a list of entities and the relationships’ between them. Format as: [Entity A] — [Action/Verb] –> [Entity B]. If you cannot identify a clear relationship or specific methodology, note ‘Ambiguous Connection.’ List unique concepts or proprietary terms.”
• Purchase Intent. “Act as a corporate buyer evaluating consultants/services. Based only on the text provided, answer the following questions. If the text does not provide the answer, state ‘Not Found.’”

1. What is the specific economic outcome or ROI in numbers/dollars?
2. What is the proprietary method or framework used to achieve this?
3. Who are the specific partners or verified authorities that vouch for this service?
4. Why should I choose this over a generic competitor?

• Differentiation Analysis. “I am providing a description of a client’s mission. Does it identify a specific problem that others ignore?”
• Semantic Density. “Identify core topics in the text. List the specific technical terms, industry standards, or locations mentioned. Suggest 3-5 authoritative nouns or proper nouns that would increase the text’s authority rating.”

GEO work may prove harder to predict and attribute than algorithmic SEO. Still, it opens new opportunities for brand marketing specialists to raise the profile of companies and influencers than in ranked results. Winning this game takes a purpose-driven, inherently differentiated content strategy–the single best way to demonstrate value to both humans and AI crawlers.

The Essential Purpose Branding Library

• ADAPT: Scaling Purpose in a Divisive World
• Purpose Campaign Trends: Anchoring Brands in a Fast-Moving World
• The Purpose Report: Corporate Purpose, People, Product and the Will to Adapt