HOW LLMS ARE CHANGING SEO & MARKETING STRATEGY

Posted on by Andrew Palios

Over the past year, generative AI tools (Large Language Models, or LLMs) have begun to transform how consumers search for information online. Traditional SEO, aimed at ranking high in Google’s or Bing’s results, is no longer the only game in town. Consumers are rapidly adopting AI chat platforms (e.g. ChatGPT, Google’s Gemini, Bing Chat, Perplexity) for search queries and recommendations. For example, a late-2024 survey found 58% of consumers had used GenAI tools for product/service recommendations (up from 25% the year prior), and one study noted a 1,300% surge in AI-driven search referrals to retail sites during the 2024 holiday season (hbr.org). At the same time, the rise of “zero-click” searches, where AI-powered summaries on search pages answer users’ questions directly, has cut into website traffic. About 80% of users now rely on AI summary answers in at least 40% of their searches, contributing to a 15–25% drop in organic clicks on average (bain.com). As IDC observes, “for nearly two decades, SEO dictated how brands achieved visibility online… That world is changing” with AI delivering answers directly (idc.com). This shift introduces major differences in how search works and forces marketers to rethink their strategies for the future.

Traditional SEO vs. LLM-Based Search: Key Differences

  • Search Results vs. Answers: In traditional search, a user types a query and gets a ranked list of links. SEO’s goal was to get your webpage as one of those top blue links. By contrast, an LLM-driven engine (like an AI chatbot or Google’s AI summary) doesn’t simply list websites, it synthesizes a single answer or a concise summary drawn from many sources. In other words, LLMs “don’t rank pages; they synthesize responses and recommend options,” leaving “fewer opportunities for discovery” and much higher stakes for being included in that one answer (idc.com). Users often take the AI’s answer at face value, rather than clicking through multiple sites. (Studies show only ~8% of people bother to click the source links in Google’s AI summaries (properexpression.com). This means visibility in an AI-generated answer is more critical than ever. If your brand isn’t mentioned by the LLM, the user might never even see your website.
  • Longer, Conversational Queries: People interact with LLMs in a more natural, detailed way than with search engines. Traditional search queries tend to be short (2–5 keywords) because users learned to feed search algorithms the minimal terms needed. LLM searches, on the other hand, “thrive on detailed, context-rich prompts” (properexpression.com). Users are asking full questions or describing their situation in sentences. In fact, the average query length has grown significantly. Search terms of 7–8 words have nearly doubled since ChatGPT’s launch (properexpression.com). Because LLMs can handle nuance, a query might be “Who are the best sustainable fashion brands that offer free shipping?” instead of just “sustainable fashion brands.” The AI will then tailor its answer based on that context (e.g. filtering by sustainability and shipping policy). Moreover, search is becoming a multi-turn conversation: with a chatbot, users often follow up, ask for clarification or personalization, and the AI remembers context from earlier in the chat. Unlike a one-and-done Google query, an LLM can act like a virtual assistant in an ongoing dialogue (properexpression.com). From a marketing perspective, this feels closer to a personal recommendation or word-of-mouth conversation than a directory of links. It creates new opportunities to build trust by being part of the AI’s suggested answers throughout a customer’s decision journey rather than just appearing once on a SERP.
  • Ranking Signals – Links vs. Mentions: Traditional SEO operates on algorithms (like Google’s PageRank) that heavily weigh backlinks, keywords, and other on-page factors to decide rankings. Essentially, the “currency” of old search was getting reputable sites to link to you and using the right keywords. LLMs work very differently. As SEO expert Rand Fishkin explains, “the currency of large language models is not links… The currency of LLMs is mentions (specifically, words that appear frequently near each other) across the training data.” (sparktoro.com) In simpler terms, an AI like ChatGPT isn’t tallying backlinks; it’s drawing on patterns in its training corpus. If an LLM has seen your brand or product mentioned often (and in the right contexts) in its training data or live index, it’s more likely to include you in its answer. This has huge implications for marketers: success in the LLM era is less about optimizing a single page’s SEO juice, and more about ensuring your brand is widely mentioned and associated with relevant topics across the web. It’s a bit like SEO meets PR. The more authoritative context your brand appears in (news articles, expert roundups, Q&A pages, etc.), the more an AI might “think” to recommend you when answering a user’s question.
  • Granular Content Retrieval: Another difference is that LLM-driven search can pull from any part of the web, not just the top 10 blue links. Google’s search might ignore a page that isn’t deemed rank-worthy for a query, but an LLM could still surface a useful snippet from that page if it addresses the question well. Generative AI can leverage small chunks of content from deep within articles or databases (properexpression.com). This both “opens the playing field” and increases complexity for SEO: even pages that would never rank highly on Google might contribute to an AI’s answer if they contain a relevant fact or phrasing (properexpression.com). In practice, it means every piece of content on your site (and about your site) is potentially important. LLMs might quote a single sentence buried in your blog post or a user review about your product. Marketers must therefore pay attention to content depth and clarity throughout their site (and beyond), not just the pages they think will rank.
  • Fewer Clicks, Different User Behavior: Because AI provides the answers up front, users are clicking fewer results. This zero-click phenomenon isn’t entirely new (featured snippets on Google have done something similar), but AI takes it further. A user might ask a chatbot for “best budget smartphones” and get a short list of models with summaries, without ever visiting a tech review website. This means traditional web traffic and click-through rates from search can decline. (Publishers have already reported traffic drops directly due to AI answer boxes, some sites saw organic traffic fall 15–30%, and in extreme cases up to ~50% after AI content appeared (bain.com) (searchenginejournal.com.) On the flip side, when users do engage, they may be highly qualified leads. Someone who bothers to click through an AI result likely has serious intent. In fact, some marketers note that while AI search may drive fewer clicks overall, those clicks can be higher-intent (closer to conversion) because the AI has pre-filtered information for relevance (properexpression.com).

Impact on Marketing Strategy and the Future of Marketing

The emergence of LLM-based search is fundamentally reshaping the marketing playbook. Here are some key impacts and shifts for marketers:

  • Loss of Control Over Visibility: In the past, if you mastered SEO you could reliably drive traffic to your site. Now, even a #1 Google ranking might get buried below an AI-generated summary. For instance, Google’s new Generative Answer Overviews appear at the top of results and can satisfy the query without the user scrolling to organic links (idc.com). This means brands can no longer count on even perfectly optimized content being seen. Even more starkly, when users go directly to chatbots like ChatGPT or Perplexity (bypassing search engines entirely), your organic ranking doesn’t matter at all. If the AI’s knowledge or its chosen sources don’t include your brand, you’re invisible in that exchange (idc.com). As IDC puts it, “a page may be perfectly optimized for keywords yet never shape how an AI model recommends a brand” (idc.com). Marketers face the reality that SEO alone isn’t enough to guarantee visibility in an AI-driven world.
  • High Stakes of Inclusion (or Omission): Marketing leaders are warning that if “your brand is absent or misrepresented” in LLM-powered search results, “you won’t even be on customers’ radar.” (idc.com) In AI chat responses, usually only a few options or names might be recommended. Being one of those few has huge value and being left out could mean losing entire segments of customers. This raises the stakes for brand representation in AI. It’s not just about ranking #5 vs #2 on a page; it’s about the binary of in the answer or not. Additionally, if the AI’s information about your brand is inaccurate or outdated, that misinformation might be amplified to countless users, harming your reputation (idc.com). (We’ve already seen chatbots confidently spout false claims about people and companies.) Marketers must now monitor not only what is being said about their brand on the web, but also what an AI might infer or generate about their brand.
  • Reduced Traffic, New Customer Journey: As noted, many users are getting their answers without clicking through. Bain & Company found 60%+ of searches now end without a click, even among users skeptical of AI (bain.com). This “zero-click” trend cuts off the traditional marketing funnel entry point. Fewer visitors land on your homepage or blog via search. This forces marketers to find new ways to engage audiences upstream, potentially within the AI answers themselves. The customer journey is becoming “an algorithm-driven narrative” (bain.com), where an AI might guide a user from a broad query, to a refined idea of what they want, to suggesting a specific product all without the user reading multiple websites or seeing the brand’s own site until maybe the final purchase step. Marketers need to ensure their brand is woven into that AI-guided journey (for example, being one of the options the chatbot recommends when the user is narrowing choices).
  • Examples of Disruption: We’re already seeing business impacts from these changes. One striking example: the education tech company Chegg saw its web traffic and stock price nosedive after Google introduced AI-generated answers that essentially used Chegg’s content to answer homework questions. Chegg reported a 49% drop in search traffic year-over-year and a 24% revenue decline in late 2024; its market cap plunged 98% (from $17B to under $200M) (searchenginejournal.com). Chegg even filed a lawsuit accusing Google of using Chegg’s proprietary content to feed its AI, siphoning away users who no longer needed to visit Chegg’s site (searchenginejournal.com). Another example: media publishers like Penske Media (which owns Rolling Stone, Variety, etc.) have sued over AI summaries cutting into their traffic and ad revenue, noting that 20% of searches for their content now show AI overviews, causing click-through rates to fall and affiliate revenue to drop by one-third (searchenginejournal.com). These cases illustrate the immediate threat to businesses: if your content is being absorbed and delivered by AI, your site may not get the visit, credit, or conversion. Marketers in content-driven industries especially need to grapple with this new reality.
  • Shifting Trust and Brand Discovery: On the other hand, being featured in an AI’s answer can be a brand boom. Users often trust the AI’s recommendations, sometimes even more than a generic search result. (Adobe found 77% of ChatGPT users use it like a search engine, and 30% trust it more than traditional search (properexpression.com). If an AI assistant consistently mentions your brand as a top solution, it can build implicit trust. In fact, early evidence suggests that when a brand appears in AI-generated summaries, it boosts brand recognition and even user response to ads. One study noted that brands cited in Google’s AI snapshot saw a 39% higher click-through rate on their subsequent ads, likely because users had just “heard” of them via the AI and assigned them more authority (properexpression.com). In a sense, getting a favorable mention from an AI is like word-of-mouth marketing on steroids. The AI is vouching for you to potentially millions of users. This dynamic will play a growing role in how brands cultivate trust and awareness.
  • New Competitive Landscape: As AI search integrates into consumer behavior, there’s a first-mover advantage for brands that optimize early. Just as early adopters of SEO in the 2000s won disproportionate traffic, today the “first movers” figuring out LLM optimization are capturing outsized share of AI recommendations (idc.com). Competitors who lag may find themselves unseen in a few years. Furthermore, customer expectations are evolving: users are starting to ask AI for highly specific or personalized recommendations (e.g. “Which skincare brands recommended by dermatologists are cruelty-free?”). If the AI has been trained on content that establishes your brand as meeting those criteria, you stand to gain; if not, you’re filtered out in the blink of an eye. Marketers will need to think about brand attributes and content (sustainability, local sourcing, expertise, etc.) that AIs might use as filters (idc.com). In short, marketing strategy is shifting from just persuading human customers, to also “persuading” or satisfying the AI algorithms that act as gatekeepers in the discovery process.

Given these changes, how can brands respond? Below I outline how marketing and SEO practices are evolving and what concrete steps brands should take to remain competitive in this AI-driven search landscape.

How Brands Can Optimize for an LLM-Driven Search Landscape

To succeed in the new paradigm, companies must update their playbook. Experts recommend a mix of technical adjustments, content strategy shifts, and even mindset changes. Here are key strategies, drawn from recent articles and case studies (2023–2025), on how to optimize for LLM-based search:

  • Audit Your Brand’s AI Visibility: Just as you might audit your search rankings, start by checking how (or if) your brand appears in AI-generated answers. “Audit your brand presence in LLM systems,” advises IDC, if you search for your brand or product in tools like ChatGPT (with browsing), Bing Chat, Perplexity, etc., do you show up? In what context? This defines your new baseline; if any important consumer query about your industry yields AI answers with no mention of you, that’s a red flag (idc.com). Treat it like a visibility gap to close. Some companies are now using AI visibility tracking tools (e.g. Peec.ai, Profound, etc.) to measure metrics like “how often our brand is cited by AI in response to X topic” (mintcopywritingstudios.com) (neilpatel.com). The first step is knowing where you stand: identify the high-value search topics or questions in your niche, and see what the AI is recommending. If it’s not you, note which competitors or reference sites are showing up instead.
  • Ensure Crawlability and Structured Data: Technical SEO fundamentals become even more important for AI. LLMs often rely on search indices (for instance, ChatGPT’s browsing uses Bing’s index, and Google’s Bard/Gemini uses Google’s) to fetch current info. So all the old basics, having a well-structured, fast site with clear sitemaps, schema markup, and no walled-off content still apply. Bain’s researchers emphasize optimizing for AI crawlability: content should be easily parsed by AI, which means using semantic HTML, adding structured data (schema for products, reviews, FAQs, etc.), and avoiding formats that AIs can’t read (e.g. text buried in PDFs or behind logins) (bain.com). Consistency is also crucial: in local SEO context, inconsistent name/address info can confuse AI models and make them less confident about a business (neilpatel.com). Likewise, using schema markup to clearly label things like organization info, product attributes, and FAQ answers helps ensure the AI correctly understands facts about your brand. In short, make your site machine-friendly much like traditional SEO, but with an even stronger emphasis on structured, unambiguous data that an LLM can ingest.
  • Create AI-Ready Content (Clarity, Context & Authority): Content strategy for LLMs revolves around producing high-quality, authoritative material that can be easily digested and reused by AI. This means writing in a very clear, factual style (the AI will be more likely to quote or rely on text that reads as authoritative and straightforward). It also means structuring content in small, meaningful chunks. For example, using descriptive headings, bullet lists, and Q&A formats that an AI can snippet-ize. One effective format is FAQ pages or Q&A sections within articles, directly answering common questions about your domain. Marketers have found that FAQ-style content “translates extremely well into AI-generated answers,” since LLMs excel at pulling concise question-and-answer pairs (neilpatel.com). Comparison guides, how-to explainers, and detailed definitions are also useful formats, because they break down complex ideas in ways an AI can easily follow (neilpatel.com). Essentially, you want to anticipate the questions users might ask the AI, and ensure your content directly answers those questions (in natural language). This increases the odds that the LLM will incorporate your text when formulating a response.
  • Emphasize Depth and Expertise: LLMs, being trained on vast swaths of internet text, look for signals of expertise and depth. In traditional SEO you might avoid highly technical jargon to appeal to general users, but with AI it can help to demonstrate subject-matter depth. One guide notes that while SEO content often simplifies terminology, “LLMs will see complex terminology as a signal of authority and depth of content.” (properexpression.com) Don’t be afraid to cover niche subtopics or use industry-specific terms (with explanations) in your content. An AI might actually prefer an in-depth source over an overly simplified one for certain queries. The concept of “topical authority” is key: if your site has rich content on a breadth of topics in your field, AI systems are more likely to view it (and you) as authoritative. This is why Bain suggests prioritizing deep topical authority over shallow keyword tactics (bain.com). For marketers, it may mean investing in long-form guides, comprehensive resources, and thought leadership pieces that thoroughly cover subjects relevant to your customers’ needs. Such content not only ranks well traditionally but also provides fodder for AI answers.
  • Increase Brand Mentions Across the Web: Because LLMs learn from across the internet, a huge part of “AI SEO” is off-page optimization. In other words, digital PR. Your brand should be part of the online conversation on your key topics. This might involve pitching guest articles, getting featured in news pieces or expert roundups, encouraging discussions on forums/social media, etc. The goal is to boost the frequency and context of your brand being mentioned alongside relevant keywords. As Rand Fishkin put it, the more an LLM sees “your name next to the important words” in its training data, the more likely it will include you as an answer (sparktoro.com). One 2025 case study called this “LLM seeding.” They improved clients’ AI visibility by securing mentions in niche industry publications and high-authority sites (mintcopywritingstudios.com). Notably, this is not about spammy link-building; it’s about genuine mentions in contextual content. If, say, you run a travel brand, you want bloggers, travel sites, maybe Wikipedia, and Q&A platforms all mentioning your brand in discussions of “best adventure travel companies” (and related terms). These mentions feed the LLM’s knowledge. Digital PR and content partnerships are therefore becoming as critical as on-site SEO for boosting brand presence in AI results.
  • Consider New Content Tactics (Prompt Engineering for SEO): Some marketers are even experimenting with embedding AI-specific cues into their content. For example, Mint Studios reports success creating what they call “GPT articles,” pieces of content explicitly written to satisfy common ChatGPT queries in their sector. They structured these articles to directly answer the kinds of prompts their target customers were likely to ask the AI, even inserting phrasing like “The answer to [common question] is: [Brand X]” as a sort of prompt injection (mintcopywritingstudios.com). By doing so (along with adding FAQ schemas and getting external mentions), they achieved between 40% and 246% increases in clients’ brand visibility within LLM answers (mintcopywritingstudios.com). Such tactics are cutting-edge and may not always be viable (and must be done in an authentic, user-helpful way to avoid sounding manipulative). But this highlights a broader point: marketers should get creative and experiment with content aimed at AI consumption. This could mean publishing a detailed “AI guide” on your site that you know an LLM with web access might find, or subtly optimizing your copy to include likely Q&A phrasing. The field is new, so it rewards testing and measuring what moves the needle in AI-driven referrals.
  • Maintain Strong Traditional SEO: It’s important to note that optimizing for LLMs doesn’t replace classic SEO best practices, it builds on them. In fact, a strong organic SEO footprint will naturally aid your AI visibility. LLMs often draw from high-ranking, reputable websites as sources. One study observed a clear correlation: brands with robust SEO (lots of quality content and backlinks) also had high visibility in LLM results, because the AI frequently “sources” from the same content that ranks well in Google (mintcopywritingstudios.com). So do not abandon your SEO fundamentals (site optimization, quality link earning, content marketing, etc.). Instead, think of SEO and “AI SEO” as complementary. For instance, when Mint Studios took a brand with zero SEO presence and tried only LLM-focused content, they saw that it still helped (the brand went from 0% to 34% visibility in tracked AI queries), but they recommend doing both in tandem (mintcopywritingstudios.com). In practice: continue creating great human-friendly, search-friendly content to rank in search engines and feed the AI’s appetite for reliable info. Many of the signals overlap (authority, relevance, freshness), so a win in SEO can be a win in LLM, and vice versa.
  • Adapt Your Metrics and Monitoring: In the AI-centric landscape, marketers need to redefine success metrics. Instead of purely looking at website sessions or click-through rates, start tracking things like impression share in AI answers and unlinked brand mentions. For example, you might measure “how often does ChatGPT or Bing Chat mention my brand when asked about [product type] in the last month?” Even if no click occurs, that exposure has marketing value (brand awareness, influence on consideration). Bain suggests shifting focus to “search impressions and AI reach” rather than just clicks (bain.com). Likewise, SEO experts advise tracking metrics such as referral traffic from AI tools, the volume of brand mentions in AI outputs, and branded search trends as proxies for your visibility (neilpatel.com). New tools and features are emerging for this (e.g. Google Search Console now shows if your pages were used in Google’s AI snapshot). Make sure to collect this data and incorporate it into your KPIs. This will also help prove ROI as you invest in AI optimization. Case in point: some brands have begun monitoring how AI-driven recommendations translate to downstream direct traffic or conversions (for instance, seeing a spike in direct visits or brand searches after being named by an AI). Marketers should broaden their analytics to capture these indirect effects.
  • Protect and Curate Your Data: With AI models training on public data, consider what information about your brand is out there. Ensure your official facts (founder names, product specs, store locations, etc.) are up to date on sources an AI is likely to scrape e.g. your website’s about page, Wikidata/Wikipedia, Google My Business, etc. This can mitigate incorrect details propagating. Also, be cautious with what content you allow to be indexed or used by AI if it might undermine you (some companies are now blocking certain pages from AI scraping to prevent, say, premium content being given away by an AI). It’s a balancing act between openness (so AIs can learn about you) and strategic control.
  • Invest in AI Expertise and Experimentation: Finally, organizations should treat LLM optimization as a new discipline, not just a tweak to existing SEO. As IDC emphasizes, “LLM optimization is not SEO by another name” and requires new skills and mindset on the marketing team (idc.com). This might involve training your SEO/content teams on AI technologies, hiring experts in data analytics or NLP, and encouraging R&D. Set aside budget to experiment with emerging platforms (for example, figuring out how to get your products listed in AI-driven shopping assistants or how voice assistants like Siri/Cortana’s evolving AI capabilities might surface your brand). Early adopters who learn what works in this space will have an edge. In fact, analysts predict that by 2029, companies will spend 5× more on LLM optimization than on traditional SEO (idc.com), underscoring how central this will become to marketing. Marketing leaders should start planning for that shift now, allocating resources to AI search strategy so they aren’t left behind as the technology matures.

Conclusion

Traditional SEO isn’t dead, but the rules of digital marketing are undeniably being rewritten by LLMs and AI-driven search. In the future, a brand’s success may hinge on whether an algorithmic assistant trusts and knows that brand enough to recommend it to consumers. Marketing is thus moving from just competing for search engine rankings to competing for AI relevance. Brands that adapt by ensuring they’re visible, credible, and relevant to these AI systems will be the ones discovered and recommended in the new era. Those that don’t risk becoming invisible as customer discovery shifts to AI-curated channels.

The broad consensus of recent articles is that marketers must act boldly and proactively: double down on high-quality content and data consistency, engage in new tactics to influence AI outputs, and measure what matters in an AI-first world. As one Harvard Business Review piece succinctly put it, the companies that “forget what you know about search” and optimize for LLMs will be poised to thrive (hbr.org). In sum, the rise of generative AI is not the end of marketing; it’s a new beginning, and it’s those willing to innovate and experiment who will write the next playbook for digital marketing success.

Sources

Unknown's avatar

Published by Andrew Palios