About us
GEO vs SEO Explainer for B2B Marketers Featuring Aviv Shamny
Limy AI’s CEO, Aviv Shamny, explains why traditional B2B SEO fails in AI search (ChatGPT, Perplexity, Gemini) and what brands must do to be found.
If you’ve been banking on your existing SEO strategy to carry you through the AI revolution, Aviv Shamny, CEO and Co-Founder of Limy AI, has some tough news: it won’t work. While Google isn’t disappearing tomorrow, the way people search for information has fundamentally changed. Your optimization strategy needs to change, too.
Aviv offers a fresh perspective. AI agent behavior is actually more important than what humans are doing on your website.
AI agents don’t behave like humans and why it matters
Traditional SEO relies heavily on clickstream data: tracking where humans click, what they engage with, and how they navigate websites. But AI agents don’t behave this way at all.
AI agents don’t even visit your client side. They actually go through the pipes of the website, which could be any CDN layer that connects between the internet and the brand.
Instead of following human navigation patterns, AI agents work in vector space search. They tokenize words, convert them into vectors, and look for similarities. Essentially they are trying to hunt for content that best matches the intent behind a user’s prompt.
This means AI agents bypass your carefully crafted user experience entirely. They’re not impressed by your conversion-optimized landing pages or your intuitive site architecture. They’re looking through the technical infrastructure of your website for semantic relevance.
The shift from algorithm-based ranking (Google) to vector-based retrieval (AI search) changes everything about how you should approach optimization.
Forget the SEO traffic acquisition craze
One of the most jarring changes marketers are experiencing is a dramatic decrease in website traffic. But Aviv argues this isn’t necessarily bad news—if you know how to interpret it correctly.
It might have decreased from 100,000 visits a month to 40,000 visits a month. And these 40,000 will probably be much more engaged with your website. They’ll know what they’re looking for.
This shift is happening because AI engines want to keep users within their platforms. Unlike Google, which monetized by sending people to websites as quickly as possible, ChatGPT, Perplexity, and Gemini want users to complete their entire research journey before leaving.
This means when someone finally does visit your website from an AI search, they’ve already:
- Asked multiple questions
- Compared alternatives
- Refined their requirements
- Identified what they want
They’re not browsing. They’re much more ready to buy.
Aviv recommends marketers shift their measurement mindset from pure traffic volume to visitor-to-conversion ratios. A higher conversion rate on lower traffic often indicates better targeting and more qualified leads.
The trick is catching and interpreting Agent behavior
Aviv suggests marketers start paying attention to which AI agents are visiting their websites and, more importantly, which URLs they’re crawling.
A practical example to illustrate this concept: If you run an e-commerce site selling running shoes and notice a spike in agent traffic to your running shoe category pages, that’s a strong signal that people are asking AI tools about different kinds of running shoes and your site is being used as a source.
This agent behavior data can reveal the topics your audience is actively researching and which content is most relevant. If your content is consistently referenced, this is also a great signal that your brand is seen as authoritative for that topic.
But here’s the catch: most analytics platforms don’t segment agent traffic effectively. This is where specialized tools like Limy AI come in—integrating at the CDN level to track agent behavior and decode what prompts are driving those visits.
How Limy AI does it: CDN-level analysis
One of Limy AI’s key innovations is integrating at the CDN level to monitor and influence how AI agents interact with websites. This technical positioning allows them to:
- See agent behavior in real-time at the server side
- Identify which vector spaces trigger agent visits
- Reverse-engineer the prompts driving traffic
- Optimize content for semantic relevance
For marketers without access to these specialized tools, Aviv still recommends analyzing your server logs to identify agent traffic patterns and the specific URLs they’re accessing. Check out our article on tracking agent behavior through server logs for a starting point.
When it comes to content strategy, think topic clusters
Showing up once in an AI search result isn’t enough. You need to own the entire conversation thread.
If ChatGPT recommends your running shoes when someone asks “What are the best running shoes?” that’s great. But what happens when they follow up with “Are they good if my knee hurts?”
I would want ChatGPT to use my website again. I would want that piece of content to be there. I would want references from third party sources… and to really own the answer and not just show up.
This requires:
- Comprehensive content coverage that addresses every question in the buyer journey
- Third-party validation from Reddit, LinkedIn, industry forums, and review sites
- Semantic clustering of related topics so AI models recognize you as an authority
You can’t just rank for one keyword anymore. You need to dominate the entire topic cluster so AI agents consistently return to your content as the conversation deepens.
Build trust where AI agents are watching
Aviv emphasizes that third-party sources matter more than ever in the age of AI search. But the strategy has evolved beyond simply buying backlinks.
Marketers need to actively participate in communities where their audience already hangs out, and where AI agents are scraping for information. Marketers should be more comfortable building a community and really showing value and making sure that the value is being translated also by their audience. And that could be on Reddit threads, Quora questions, or anywhere around the more public sources being used within their industry.
This means encouraging customers to share experiences on Reddit and review sites, participating authentically in industry forums, making it easy for satisfied customers to leave detailed reviews, and monitoring which third-party sources AI models cite in your category.
The goal isn’t manipulation. The goal is ensuring your brand shows up in the right way across the diverse sources AI models trust.
Don’t wait to act. The future is already here.
Aviv’s parting advice for marketers on the fence about generative engine optimization? The change is already happening.
As time evolves, we see ChatGPT adding more application layers and these application layers are becoming a very dominant part of our life and how we’re getting information, consuming information.
Rather than chasing vanity metrics like 500,000 daily visitors with a 0.01% conversion rate, marketers need to embrace targeted campaigns that attract the right audience with dramatically better conversion ratios.
The age of spray-and-pray SEO is ending. The age of semantic optimization for AI agents has begun. The only question is whether you’ll adapt in time to maintain visibility—or watch your competitors own the AI search conversation in your category.
Want to learn more about optimizing for AI search? Connect with Aviv Shamny on LinkedIn or try Limy AI free for 14 days at limy.ai.
