Being a web marketer today is challenging due to rapid shifts in how web content is discovered and consumed. Traditionally, organic traffic came from users searching on major engines like Google, reviewing top results, and clicking the best fit. However, AI conversational engines are transforming this pattern quickly, accelerating the change. Consequently, direct website traffic is declining year over year, paid ads are costlier with lower conversion rates, and Google’s Gemini AI-assisted results dominate searches. About 60% of searches last year ended without a clickthrough, a trend rising steadily. The value of top search rankings has dropped to less than half of what it was a few years ago. So, what’s the path forward? Enter the GEO (Generative Engine Optimization). Users increasingly don’t see links and, more importantly, traditional search engines were designed to deliver answers, not pages. AI-driven tools like Claude or ChatGPT focus on providing answers rather than directing users to links. Google’s recent “AI Mode” signals its shift in this same direction. For example, during a used car search with my daughter, almost all inquiries about features and reliability were answered directly by Claude, without needing to visit multiple sites. While not perfect, LLMs increasingly guide the customer journey outside traditional brand encounters. This raises a critical question: how can marketers influence AI-generated recommendations? The good news is much of what creates authoritative content for search engines still matters: crafting quality content, ensuring accessibility, and supplying strong metadata. Search engines—whether traditional or AI-enhanced—require context and structure to interpret content. On top of traditional SEO concerns, marketers must now focus on how AI perceives and indexes their content—a rapidly evolving discipline broadly called GEO. Understanding Content for LLMs While content valuable to humans also aids LLMs, the crucial difference lies in reasoning capacity.

LLMs process language patterns but lack true understanding. For example, humans know “AWD” means all-wheel drive and “manual” refers to a transmission type. LLMs, without explicit context, see just words—they might mistake “manual” for a book or not grasp “AWD. ” Although training data provides some context, marketers can assist by including detailed explanations on sites, such as extensive glossaries—a practice advocated by experts like Corey Vilhauer. Well-crafted FAQs also play a renewed role, focusing not just on presumed user questions but on those an LLM might ask. Providing precise answers to these anticipated queries increases the chance AI will present them. Enhancing Metadata’s Role Schema markup and embedded semantics like microdata or JSON-LD are not new, but now they have heightened importance. Schema. org offers hundreds of schemas for various uses, and while previously marketers targeted those that aid Google’s Rich Results, AI’s broader comprehension allows leveraging a wider variety of schemas that closely match actual content. A novel emerging tool is the llms. txt file, akin to sitemaps. txt but designed specifically for LLMs. Written in Markdown, this file offers an LLM a guided tour of your site, highlighting key sections, jargon, and features to improve AI comprehension. Building Authority Through Expertise Establishing authority remains critical—beyond good content, it involves visibility through collaborations, media presence, accurate Wikipedia entries, and high-quality backlinks. A site’s authority grows when others reference its content, increasing the likelihood AI cites it. Authority depends on human endorsement as much as on content quality. Old Principles, Renewed Importance Many strategies from traditional SEO remain essential but have gained urgency since content now competes for attention beyond direct website visits. Your message, site structure, and data quality must all be exceptional. There are no shortcuts, but in truth, there never have been.