LLM Marketing

Hottest topic in my limited marketing circles is how do you get your solution or company to come up as part of an answer with LLM's like ChatGPT, Claude, etc.? I have heard thoughts as to the content on websites needing to be more of a narrative for the LLM's to learn from, but that is about it. What are you hearing these days?

 

I'm hearing from many that if you are using LLM for company efficency, that you need to keep a human in the mix in order to monitor voice accuracy and quality control. Seems like a logical idea to me.

 

EEAT. I read something a few months ago about adding a LLM.txt at the root of a website to bait or attract AI robots. There's 0 proof this works.

I might be wrong here, but think about something simple : where do AI scrappers get their data ? Are websites indexable by AI robots or does the scraping technology depend on search engines ? If they indeed depend on search engines like Bing, Google, etc... how do they select their sources ?

SEO, is still the springboard in the end. EEAT + how to's + FAQ + entity-focused content are the best leads, in my humble opinion.

I don't trust the GEO, LLM consultants or whatever, who are all day and night on linkedin. No one knows the perfect method for GEO and LLM. The same truth applies for Google's algorithm. We just need to keep trying, because I also think that niches are important here. I dare anyone on Linkedin to prove me that he can make any site, in an ultra high competitive niche, to appear as recommendations on chatgpt or whatever. It's impossible.

 

@EAsports your points are all very valid. For a long time now, I have been telling anyone who will listen that it is a fool's game to base a big chunk of one's business on SEO. It is too volatile and that is by design. That said, when search engines first came out, you could "find the angles" in the then immature technology and sometimes do quite well. I am not talking about black hat, but just wondering if these newborn LLM's have any growing pains when it comes to developing their answers and if an online presence might be optimized to benefit from any such quirks.

 

@jrbiz
Hi. I humbly think SEO is still useful, but you are right about not betting everything on it, because of its time consuming nature. The thing with AEO, GEO and LLM is that all of them depends on quality content, which is a branch of SEO itself. No matter what some consultants are saying, there are absolutely no proven cheat codes or shortcut to make one business appear there.

If Answer engine optimization depends on, let's say, JSON-LD FAQ and how's to, I suppose that it follows more or less the same logic as Google's position 0. This was an eldorado a few years ago. Everyone wanted their site to appear on position 0 and before that, infography was the trend. You were not cool if you weren't following the trend. What happened some time later ? Google gave less importance to infography and pplz started to realize that being on position 0 made their CTR fall, because users wouldn't click on their pages, having found what they were looking for in the SERP.

Anyway, I'll do my best to provide quality content with the user-satisfaction as my top priority. If I'm really helping people, Google will rank me, AI will notice me (following Google's standards) and that's it. I truly believe that all the bla bla around the new AI trends are nothing more than the unmovable basics = provide quality.

 

I was one of those who did indeed experience growing pains until I purchased your Novocaine For Beginners Model tinfoil hat. Now that I have said hat....no worries.

 

I think the safest approach is to treat LLM visibility as an extension of strong SEO rather than a totally separate trick. If a business has clear entity signals, useful content, structured data, author details, and mentions from trusted sources, it gives AI systems more reasons to understand and reference it.

I would not rely too much on new hacks like special text files until there is stronger proof. For now, I would focus on pages that answer real questions in depth, show experience, and make the company easy to verify across the web.

The main goal should be becoming a source that both people and search systems can trust.

 

jrbiz, the framing that's missing in this thread is that there are two different LLM-visibility games going on, and most advice conflates them.
The first is training-data inclusion (was your site scraped before the model's cutoff). That's mostly out of your hands and rarely affects today's answers. The second, and the one that actually drives current citations, is real-time RAG. ChatGPT browsing pulls from Bing's index. Microsoft Copilot is also Bing. Perplexity uses Google plus Bing plus its own crawler. Claude's web search uses Brave's index. So "getting cited in AI answers" in 2026 is overwhelmingly an exercise in ranking well in those specific indexes for the exact phrasing the model needs to quote, not making your content "more narrative" for some imagined training pass.
The angle nobody in this thread has hit yet: GA4 already shows you which AI tools are referring traffic. Filter the Acquisition report by source, look for chatgpt.com, perplexity.ai, copilot.microsoft.com, claude.ai. The pages those sessions land on tell you which of your URLs the models are already citing. From there it becomes ordinary content optimization on the pages that already have a foothold, plus aggressive Bing Webmaster submission for new pages. EAsports is right that SEO is still the springboard, but the springboard has multiple landing pads now, and the ones that aren't Google are quietly the ones growing fastest.