How Marketers can Prepare for AI Search Models

Is the era of the Google search results page fading for discovery? The data suggests a rapid shift is underway. We're seeing a seismic shift in how information is discovered - moving rapidly from keyword searches yielding lists of links towards a conversational AI delivering direct answers (ChatGPT, Perplexity, Google Gemini, etc.). This isn't a slow evolution; it's happening fast. Consider this: Adobe Analytics recently reported a 1200% YoY increase in traffic from gen AI sources to US retail websites. While this is retail data, view it as a leading indicator. This behavioral change - seeing answers, not just links - will ripple across all customer journeys, B2C and B2B alike. I've heard from multiple B2B startups that they are getting more inbound from ChatGPT than Google Search. What does this mean for the customer journey? * Discovery & Search: Buyers won't just browse websites. They'll increasingly ask AI models for comparisons, summaries, and recommendations. * From Answer to Action: This isn't just about information retrieval. With AI agents like OpenAI's Operator, Google's Project Mariner, and Amazon's "Buy with Me", we're seeing the potential for AI to move directly from discovery to research to purchase. The Implications: * For Marketers: The traditional customer journey playbook needs a significant update. How do you ensure your solution is surfaced, understood, and trusted by the AI? Getting discovered in an answer-first world requires new strategies, likely involving structured data on both 1st party and 3rd party sites taking advantage of protocols like the emerging Model Context Protocol (MCP) to effectively communicate to the models. * For Founders: There's an opportunity to build the next generation digital experience platform. We need solutions purpose-built for this new reality across Customer Experience technology categories - consider that in a post AI world the next customer may by AI, which means leveraging conversational interfaces, agent capabilities, and protocols like MCP. It isn't just about new features; it's about rethinking the entire go-to-market motion in an AI-world.

Kayak's AI Strategy Should Wake Up Every Hotel Marketer When Kayak launched Kayak.ai, it wasn’t just a side project. CEO Steve Hafner called it a “test lab for AI-first features,” but it signals something deeper: AI isn’t just assisting search, it’s transforming how travelers shop, decide, and book. In Hafner’s words, metasearch may soon become backend plumbing for tools like ChatGPT or Gemini. Why would AI pull pricing from dozens of sites when it can get structured, bookable data from one place, Kayak? That’s the future Kayak is building for: one where AI agents sit between your potential guest and your hotel. Not browsers. Not visitors. Agents. So here’s the real question for hotel marketers: Are your offers ready for an AI interpreter not a human user? When a guest tells their AI, “Find me a boutique hotel in Chicago with late checkout and no resort fees,” it’s not sifting through your website. It’s going straight to data it understands. Your job is to be in that data. Here’s how to get there: 1. Structure your content for machines. AI doesn’t “read” web pages like people. It parses structured content. Your pricing, packages, and policies need to be machine-readable. 2. Optimize for AI-driven distribution channels. Kayak.ai is just one example. There are more coming. Track where your inventory is showing up. 3. Rethink the funnel. AI agents don’t browse. They book. They won’t nurture leads. They’ll jump from intent to action. Be ready at that exact moment. 4. Train your teams now. Prompting is no longer optional. Knowing how to guide AI internally or in the market will define tomorrow’s marketing success. That’s where I come in. I run an AI Literacy/Mindset Program built specifically for hoteliers. It’s designed to transform your teams not through top-down mandates, but from the bottom-up. It's proven to gain one hour of time per day, per person. We embrace scaling through AI Ascension, a clear path from foundational literacy to real impact: First, we get your team fluent in prompting. Then, we show them where automation fits into their actual workflows. Finally, we introduce AI agents not as hype, but as usable tools for sales, ops, and marketing. This isn’t about chasing buzzwords. It’s about equipping your people to lead AI adoption from the ground floor. Because if they don’t, someone else’s agent will own your booking. Let’s make sure it’s your team that leads.

Most AI SEO advice is fluff. You should try my "surgical" approach: CONTEXT: I was a Product Manager on Google Search focusing on data inquisition and ingestion from third party providers before I left to start AthenaHQ to help marketers win on AI search. Here's what I learned analyzing citation patterns at Google and now at Athena. AI models pull from 20 to 40 sources per query. VS Humans only check the top 3 Google results, maybe more of page 1 if you're lucky. This creates a massive opportunity for your brand. Why? When someone asks ChatGPT about cybersecurity software, it might cite a security forum that ranks # 34 on Google but has deep technical discussions. That forum carries more weight in the model than a generic tech news site at position # 5. My surgical approach works like this: Take your top 20 customer questions. Run them through multiple AI models. Map every single citation. You'll discover industry specific publications that consistently get referenced but rank between positions 20 to 40 on Google. (You can do this first step in Athena btw) THESE sites are the new arbitrage opportunity. A Forbes placement will cost you $$$ and three months of Digital PR / relationship building. Compare that to these industry sites which cost between $500 to 2,000 per sponsored placement or guest post. Plus? They're way more likely to respond to your emails within 48 hours. They're hungry for expert content. Still not convinced? Forrester data shows that 90% of B2B SaaS procurement now uses AI somewhere in the evaluation process. Whether it's generating RFPs, evaluating vendor responses, or validating tools. Your prospects ARE asking AI models questions about your category before they ever hit your website. If you're not cited in those responses, you don't exist in their research phase. -- P.S. I'm Andrew Yan (ex Google search), Co-founder of AthenaHQ (YC 2025), a GEO monitoring solution and action center empowering brands like Ollie, Checkr, Motion, and OneSignal to lead the conversation in AI search results. Do you want to run this playbook? Shoot me a DM. 

Forbes estimates 60%+ of organic traffic is being effected due to the introduction of AI summaries, the adoption of LLMs, and ultimately the changes in SEO. I recently spoke at Affiliate Summit East on the emergence of AI and the impact to eCommerce. I touched on SEO and the move to SGE (search generative experience), the issues facing identity capture and noise created by bots as well as AI-generated content, and how content goes viral with changes to TikTok & Meta's algorithms. Sharing some of the learnings (and slides) on AI summaries and rankings: Impact - Google's AI summaries can take up to 3 full mobile scrolls or 2 desktop scrolls (1500 pixels) - LLM search went from 0.25% of traffic to 2.25% in <12mo - 60% of organic site traffic is impacted meaning the CTR drops aka people do not end up landing on your site. How to adapt - SEO & SGE are underpinned by the same recipe: content. - How to show up as the featured AI summary or authority? 1. Create unique content and lots of it 2. Build contextual content At Checkmate we get 1.5M+ unique visitors/week to our public-facing product pages. We also are featured on 1700 ChatGPT pages. How we were able to do that is by pulling in long-form structured content for the products we help sell as well as using LLMs to generate some unique content. For those in eCommerce I featured Stanley 1913 with what I think as a really strong product pages that rank well for SGE. If you look at any of their products they have extremely rich, unique content. They leverage product descriptions, product specs, related other products, reviews (not hidden or collapsed) & FAQs. For SGE shorter isn't better, the more unique content the generally better the indexing. Stanley 1913 also has a really strong formatting structure with clear H1, H2 tags, and containers. You have to remember if a bot can't make sense of the text then it won't surface in AI summaries. This will also be extremely important in the future of agentic commerce. 2. Building contextual content The way people are searching is moving from "black shoes" to "best shoes for running a marathon". Context is key. To be able to show up in LLMs or AI summaries, you need to associate your products within that context. 3 easy ways to do that: 1. Create and leverage a PR strategy 2. Build your own written contextual information in a blog on your website 3. Contribute/invest to review sites If you are able to both create unique content with a strong structure and build contextual content AI summaries and LLMs can be a great source of traffic. It is early days so investing in content and structure is a must. Drop me a note if you have other tips you see working or other brands doing it well!

There's a quiet revolution happening in the world of content and discovery, and we're barely talking about it. 🚨 SEO as we've known it is quickly becoming obsolete. AI tools like ChatGPT, Perplexity, Gemini, Grok, and Claude are reshaping how buyers research, especially in high-involvement e-commerce sectors like consumer tech. Andrej Karpathy recently highlighted this perfectly: "It’s 2025 and most content is still written for humans instead of LLMs. 99.9% of attention is about to be LLM attention, not human attention." Imagine a buyer exploring a high-value product—say a premium smartwatch or a flagship smartphone. Rather than sifting through dozens of links, they're now simply asking: "Which smartwatch is better for health tracking: Fitbit Sense or Apple Watch Ultra?" This isn't hypothetical; it's happening right now. AI assistants are increasingly trusted as expert advisors that simplify complex product comparisons. For brands, this changes everything: ✅ Traditional SEO and "ranking #1 on Google" is no longer enough. ✅ AI agents handle the discovery process, dramatically shortening the consumer journey. ✅ Businesses must now optimize their content not only for human eyes but also for AI interpretation. (As Erik Wikander put it beautifully) A few eye-opening shifts already happening: 📉 Review platforms like G2 and StackOverflow have lost significant traffic post-ChatGPT. (Elena Verna) 📉 Big players like HubSpot, Figma, and Canva are seeing declining organic traffic as AI directly answers user queries. (Oliver Molander) 📉StackAI now receives more inbound from ChatGPT & Perplexity than from Google (Antoni Rosinol) The implications for Ecommerce and consumer tech brands are massive: Your product content must become "AI-native," meaning it's structured clearly enough for AI tools to pull and recommend. Content needs precise differentiation and expert-level detail because the AI gatekeepers are getting smarter at recognizing genuine value vs. fluff. As Tomasz Tunguz has highlighted, expect the emergence of AIO (AI Optimization)—the next evolution of content strategy where you're optimizing not just for search engines but for a multitude of personalized AI assistants serving diverse ICPs. In short, the future belongs to those who understand how to capture the attention of AI agents first, and users second. Traditional SEO is fading. We're entering a world where personal AI agents will act as gatekeepers, curating hyper-relevant content tailored exactly to individual needs and preferences. This demands a radical reimagining of CX tech stacks, particularly around product discovery, comparison, and commerce journeys. At Swirl®, we're addressing precisely this challenge by building specialized AI Agents—transforming customer experience into dynamic, personalized, and AI-optimized experiences. If your brand is navigating this shift and looking for ways to stay ahead, let's talk. #artificialintelligence #Ecommerce #AIO #SEO

🌶️ SPICY TAKE: Websites are becoming an outdated eCommerce tool That's because #AI is shifting the web from human-centered browsing to agent-led buying (More details here: https://lnkd.in/gCxTJKSv). As we move to machine-to-machine (M2M) commerce, #CRO and #A/B testing are likely to look very different In the near future, I envision a landscape where websites are no longer the primary touch points for shoppers Instead, I expect "websites" will be rich structured data snippets, hyper-personalized to the individual user, based on past AI agent-assisted searches and buying behavior The interface will likely be through an AI app, bot or GPT tool -- not a website or landing page From this vantage point, it's tempting to think CRO is imploding from the inside out. . . But, for forward thinkers, there are definitely opportunities. #Experimenters will be needed more than ever. Just in new places Here are the top-3 ways I anticipate experimenters will be able to reposition themselves in an M2M, AI-driven world: 1️⃣ OPTIMIZE FOR AGENT-FRIENDLY DATA AND FEEDS ⚡ AI agents, GPTs, and AI shopping bots will likely become the main eCommerce tools They'll likely rely on structured, accessible, and interpretable data As a result, they'll be increasing need for experimenters to specialize in optimizing meta data, schema, and structured data for AI product catalogs EXPERIMENTATION OPPORTUNITIES: ✅ Test tagging to determine which data format is the most visible and preferred by AI agents 🙉 The KPI will be higher-rankings in AI result/recommendations 2️⃣ DESIGN & TEST AI PROMPTS ⚡ Shopping will become more prompt-based, like “find me the best wireless headphones under $200” EXPERIMENTATION OPPORTUNITIES: ✅ Test prompt responses to favorably influence AI outputs ✅ Optimize LLM interaction design to show the most compelling product attributes 🙉 The conversion goal will become seeding AI tools to recognize and prioritize the client's brands offerings, first 3. OPTIMIZE EMBEDDED INTERFACES ⚡ With shopping likely to happen inside smart AI assistants, they'll be need to focus on multi-modal optimization (image + voice + text) and new forms of engagement EXPERIMENTATION OPPORTUNITIES ✅ Think like an AI-UX designer. Determine the journey when the AI is the customer, not people ✅ Test micro-interactions inside AI chat UIs, AR overlays, or voice interfaces ✅ Create copy that persuades bots, not people with the right algorithm inputs to see how how changes in backend data or structure impact AI result visibility 🙉 Conversion metrics will likely be "SERP" rankings in engines powered by LLMs, not just traditional SEO YOUR TURN: 📣 What do you think the future of experimentation looks like in this new AI, M2M paradigm? Share your thoughts and speculations below. Will the web as we know it cease to exist? ⬇️