3 LLM SEO Strategies That Actually Work

Most marketers are still debating whether AI will kill SEO. Meanwhile, builders like Maxime Dolores are creating the infrastructure to measure – and improve – AI search visibility.

Maxime is the founder of Doppler, which he describes as "Google Search Console for LLMs." When we connected last week, I expected another theoretical discussion about AI search. Instead, I got a masterclass in how these systems actually work.

Our conversation revealed three counterintuitive strategies that challenge everything we think we know about search optimization.

Traditional SEO focuses on ranking your own pages. LLM SEO requires a different approach entirely.

"What I'm really trying to lean in with Doppler is this opportunity finder for sources," Maxime explained. "I believe right now that is the biggest leverage that you can get."

The strategy works like this:

Doppler simulates thousands of ChatGPT conversations, then analyzes which sources appear most frequently for any given topic. But instead of trying to outrank these sources, you identify where your competitors are mentioned – and where you're not.

"If I manage to get mentioned on those sources, then anytime someone's going to ask ChatGPT for one of those tools, I'm probably going to pop up at least once."

This approach sidesteps the entire question of "ranking" in AI. Instead of optimizing your own content and hoping AI finds it, you place your brand where AI already looks.

The data backs this up. Research shows the most-cited domains in AI assistants tend to be established reference sites. Getting mentioned on these sites effectively guarantees AI visibility.

The shift from keywords to prompts fundamentally changes how we approach content optimization.

"People underestimate how the prompts are written by normal users," Maxime noted. "You have such long, detailed prompts now. You're getting so much detail from every single prompt."

Traditional keyword research might target "competitive intelligence software." But LLM users ask questions like:

Maxime's approach involves building detailed user personas, then combining them with various prompts to test all permutations. This reveals the actual language patterns AI systems encounter – and optimize for.

Recent studies on ChatGPT search behavior confirm this shift. Users provide significantly more context to AI than to traditional search engines, creating opportunities for highly specific content to surface.

Most SEO tools adapted for AI only track one metric: referral traffic from AI platforms. But this misses the bigger picture.

Doppler revealed three distinct levels of AI visibility:

Level 1: Context Inclusion Your content gets pulled into the AI's context window but doesn't appear as a citation. "ChatGPT will show you maybe five citations but it actually queried like 90 or 50 pages," Maxime explained.

Level 2: Visible Citations Your content appears as a numbered source or link in the AI's response. This is what most people think of as "ranking" in AI.

Level 3: Click-throughs Users actually click from the AI response to your site.

Most brands only measure Level 3, missing the full scope of their AI presence. Doppler tracks all three by monitoring when AI crawlers access your site, correlating this with citation patterns across different platforms.

Perhaps the most surprising finding from our conversation: traditional SEO metrics barely matter for AI visibility.

Maxime shared two examples that challenge conventional wisdom. First, his own low-authority blog posts regularly appear in ChatGPT responses. Second, a three-week-old Shopify store received its first order from a ChatGPT recommendation.

"A lot of it comes from the fact that the Bing index is really different from the Google index," he explained. "It tends to suggest pages which are really sticking to the query, whereas Google will really understand the intent."

This creates an unprecedented opportunity. While established brands fight for Google rankings, new sites can capture AI visibility by creating highly specific, relevant content. Data shows 63% of websites now receive some AI traffic, regardless of their traditional authority metrics.

Based on Doppler's data and my own testing, these tactics show the most promise:

1. Strategic Source Placement Instead of building links, build relationships with sites AI already trusts. Use tools like Doppler to identify where competitors appear, then pursue guest posts, interviews, or updates to existing content.

2. Ultra-Specific Content Creation Stop creating "ultimate guides." Start creating narrow, specific content that directly answers the conversational queries users actually ask. My guide to tracking AI traffic outperforms broader SEO content specifically because it solves one precise problem.

3. Technical Optimization for AI Crawlers Ensure GPTBot, PerplexityBot, and other AI crawlers can access your content. Consider implementing llms.txt to provide explicit instructions to AI systems.

4. Conversation Mining Use customer support logs, sales calls, and user research to understand how people actually describe their problems. This conversational data becomes your "keyword" research for AI optimization.

During our conversation, Maxime revealed what's possible with server-side manipulation: "You could literally rewrite the page content when it is a visit from an AI crawler."

This type of cloaking could inject prompts telling AI to rank certain content first. But as research on retrieval-augmented generation shows, these systems will quickly evolve to detect manipulation.

The sustainable path forward isn't gaming the system – it's understanding how AI determines trust and relevance, then building content that naturally meets those criteria.

"What you're going to see more and more is basically a replication of what Google Search Console is like, but for prompts," Maxime predicted.

He believes SimilarWeb is best positioned to capture this data: "When it comes to data about search, they've always been leveraging their extension. If anyone is doing anything about prompt volume, those are the people I would be looking at."

Until then, tools like Doppler fill the gap by providing visibility into:

Industry analysis suggests this fragmentation of search across multiple AI platforms will only accelerate, making unified tracking tools essential.

The gap between talking about LLM SEO and actually implementing it remains massive. While consultants debate terminology, practitioners like Maxime are building the future of search optimization.

Start here:

As I've written before, the fundamentals of good content haven't changed. But the game has new rules, new metrics, and new opportunities for those willing to learn them.

The question isn't whether LLM SEO matters. It's whether you'll figure it out before your competitors do.

Watch my full interview with Maxime Dolores for a complete walkthrough of Doppler and more tactical insights on AI search visibility.