LLMO and GEO: What We Know About Optimizing for LLMs and AI

ChatGPT, Gemini, Claude, and Perplexity are all Large Language Models (LLMs) that offer an alternative to traditional search. The difference? It’s searching and conversing. Instead of links, it’s full answers. To compete, search engines have AI Overviews and accompanying LLMs in their experience. These are transforming the way people discover products, services, and brands.

So, how do you optimize for AI to keep up? The hard truth—everyone is still figuring this out. Be wary of claims to the contrary. LLM optimization is still an evolving space, and there’s no definitive formula.

The good news? Large Language Model Optimization (LLMO) and Generative Engine Optimization (GEO) align with traditional SEO efforts. Your SEO team should be tackling both.

Let’s take a look at what we do know, what we don’t know, and what our experts are doing about LLMO and GEO.

What is LLMO and GEO?

We’ve thrown a lot of new acronyms at you. Let’s break this down. While not apples to apples, a comparison of each reveals a similarity in goals between traditional SEO and AI tactics: visibility.

Search Acronyms:

• SEO: Traditional Search Engine Optimization (SEO) targets the search engine results pages (SERPs) of your key users by focusing on relevant keywords, securing high-quality backlinks, and optimizing page-level elements.
• Technical SEO: A part of SEO, these changes are made to a website’s schema and other backend architecture so that search engines can crawl and index your pages.
• GEO: Generative Engine Optimization focuses on content to populate generative AI platforms. Its goal is to serve your content as the source for AI-generated answers.
• LLMO: Large Language Model Optimization focuses on the AI behind chatbots and search features, ensuring your content is usable by the technology.

The goal now is to utilize both traditional search and AI search optimization tactics with a new mindset: It’s not about rankings. It’s about relevance, trust, and clarity, which puts you in both SERPs and AI results.

Why do these new-fangled acronyms matter? Buyers are increasingly bypassing traditional search journeys and making decisions based on conversational responses instead of page-one rankings.

Why LLMO and GEO matter

We’re witnessing what many call “The Great Decoupling” or the “crocodile effect.” While impressions are rising, click-through rates are plummeting, especially where Google’s AI Overviews are present. According to recent data, position-one CTR drops by over 34.5% when AI overviews are shown, with overall search clicks falling 30% year-over-year, despite impressions rising 49%.

Impressions are rising, click-through rates are plummeting.

What we are seeing:

• Click-through rates (CTR) are collapsing
• Citations are rising from deep results (items further down in SERPs)
• Query complexity is rising

This data confirms what we feel: users are searching differently, and AI is responding in kind. Let’s dig deeper.

You don’t have to rank on page one to be cited as a source within AI Overviews, meaning if your competitors are outranking you, it doesn’t mean they will show up in an AI result over you. Looking back over the trailing 12-month period, new data shows a massive increase in AI Overviews, especially in queries of eight words or more.

Interestingly, citations from outside page one results have dramatically increased YOY:

• 48% rise in technical terminology in search queries
• 400% increase in citations from results in positions 21–30
• 89% of citations come from beyond the top 100 organic listings

Consider your industry as well. Recent data may surprise you in terms of what is showing up in AI. The same data shows that only 4% of core e-commerce commercial queries include AI Overviews, while B2B sees a higher rate of 70%.

AI Overview performance by industry.

This isn’t all bad news. It presents a significant opportunity—if you know how to optimize for it. So, how should you capitalize? Hint: It’s not about gaming the algorithm—it’s about becoming the trusted answer.

What we know about winning LLM visibility

Best practice time! Based on what we know, these key strategy pillars below will position your brand to be trusted and successful within LLM queries.

What we see working in winning LLM visibility.

1. Help LLMs Understand Your Expertise

LLMs decide what to show based on clarity, structure, and thematic consistency. Your job is to help them understand what you do, how you help, and why you’re credible.

• Ensure your structured data is immaculate—Product Schema, Reviews, and Inventory status need to be flawless
• Create clear product categorization that aligns with how users naturally think about products
• Develop comprehensive content addressing multiple aspects of products for query fan-out— brand, shape/size/SKU, materials, area of industry, requirements, and type (B-SMART)

2. Extend Your Answers to Real Audiences

LLMs learn not just from your site, but from the entire web. You need to show up where the conversations are happening.

• Ensure your content addresses real questions: Especially long-tail, high-intent queries.
• Engage in communities: Participate in forums like Reddit, Quora, and industry-specific platforms or industry associations like FMA or PMA. Continue to encourage reviews and ratings.
• Maintain a publishing cadence: Continuously updated Q&A and blog formats help signal freshness and relevance.

This isn’t just a B2C play. B2B companies benefit enormously from answering niche questions and contributing expert commentary where their audiences engage.

3. Deploy a Technical Foundation That AI Can Read

Behind the scenes, technical clarity gives your content a better chance of surfacing in AI tools.

• Use schema everywhere: Define context and relationships between content types.
• Implement IndexNow: Push changes to search engines in real-time for faster reflection in AI outputs, and recommended by Microsoft for their AI experiences.
• Adopt llms.txt: A new protocol that guides LLM crawlers to your most useful, machine-readable content.
• Limit JavaScript reliance: Server-rendered content is more easily parsed by AI systems.

AI isn’t just scanning your home page—it’s dissecting the structure of your entire website. Make sure it’s understandable.

How to measure LLM success

Success isn’t just clicks — it’s presence where AI is answering questions. So just as you look at your traffic sources to help measure SEO performance, you’ll want to see increased referral traffic and leads from LLM-powered tools.

Tools and tactics you can use today:

• Referral Traffic From LLMs: Track using UTM parameters or direct sources from tools like Perplexity, ChatGPT (browser plugin), and Bing AI.
• AI Overview Tracking in SERPs: A custom method flags Google results that show AI Overviews and detects citations.
• AccuRanker & Ahrefs Integration: Flags when an AI Overview is triggered and whether your site/content is included.
• Lead Attribution Tools: Use Loop Analytics and CallRail to identify if AI-sourced traffic converts.

Future OuterBox Custom Report

The technology is still young, and so are the tracking tools that accompany it, but advanced tools are emerging rapidly. Here are additional methods you might be using in the near future:

• LLM Query Tracking: Tools are emerging to log prompts and detect brand mentions across platforms.
• Citation Monitoring: Expect API-level visibility into when/where your brand appears in AI summaries.
• Knowledge Graph Accuracy: Improved entity tracking will indicate semantic presence and brand trustworthiness in AI tools, enhancing their overall accuracy.
• Long-Term KPI Shift: From ranking to inclusion, from clicks to answer presence.

The Future of Search Starts Now

Being found in AI platforms is now essential. Whether you’re targeting consumers or businesses, your brand needs to establish its presence in traditional search rankings and AI-generated answers.

Our biggest takeaway is to stay on your toes! This space is evolving fast, and be ready to pivot. However, based on what we know now and the foundations SEO has given us, the strategies you build today create the trust signals that AI will rely on tomorrow. Start now—and stay ahead.