Optimizing for LLMs: It’s Not Just “SEO With a New Name”

Optimizing for LLMs It’s Not Just SEO

There’s a loud new wave in the SEO space. “Optimize for LLMs,” they say. “Be ready for ChatGPT search.” And while it sounds shiny and advanced, let me tell you straight ranking in LLMs is way messier than traditional SEO.
This post dives into why that is, what Daniel Foley Carter recently nailed about this chaos, and what people like us who actually do the work can do to adapt without the BS.

“We’re witnessing a mass capitalization of capitalizing on client wants without having the solution.”
Daniel Foley Carter

Daniel Foley Carter Said It Best

Daniel dropped a truth bomb on LinkedIn recently, and it hit home. The rush to package and sell “AI SEO” has led to more confusion than clarity.
Let’s be real: Ranking or being cited in LLMs isn’t something you can “optimize” for with a checklist.
These tools don’t even work like Google.

Let’s Break It Down (Facts First)

1. Every LLM Is Different

OpenAI’s models aren’t trained the same way as Google’s Gemini or Anthropic’s Claude. Each model uses unique pretraining data and logic. No single strategy fits all.

2. Training Data ≠ The Real-Time Web

Just because you publish a blog doesn’t mean GPT-4 knows about it. Many LLMs still hallucinate or infer info from partial context. Being online ≠ being understood.

3. RAG (Retrieval-Augmented Generation) Isn’t Predictable

Even if an LLM fetches content from the web, there’s no guarantee it fetches yours. Or that it fetches correctly. Or that it even shows citations.

4. LLM Memory & Bias Exist

Some platforms personalize answers over time. If a user interacts a lot, their future queries are biased. Meaning your brand might show up in one session but not the next.

5. LLMs Make Stuff Up

Feed them a file they can’t read? They’ll guess based on the filename. The outputs aren’t just unstable they can be straight-up imaginary.

What Should SEOs Actually Do?

Here’s Daniel’s advice plus my own take, as someone navigating this alongside you.

✅ 1. Learn How LLMs Work

Before you “optimize,” understand what you’re optimizing for.
Watch this LLM primer and dive into 3Blue1Brown’s neural network videos. Seriously ditch the fluff podcasts for a week.

✅ 2. Stop Worshiping Templates

LLMs aren’t like Google. SOPs won’t save you here. Your goal isn’t to control them it’s to influence them through broader signals.

✅ 3. Play the Long Game

You want to be cited by an LLM?

  • Build a brand that’s talked about
  • Publish useful, high-quality, human-first content
  • Distribute like a machine
  • Keep showing up

✅ 4. Think Past Keywords

Ranking now includes:

  • Being cited across the web
  • Mixing content formats (video, audio, text)
  • Earning mentions from credible sources

✅ 5. Focus on CRO

Clicks are shrinking. Impressions? Going up.
That means the smart move is to convert more with less. Maximize the value of the traffic you do get.

✅ 6. Distribute Like a Pro

Write that blog → make a YouTube short → chop it for Insta → link everything → embed in the article → push via email.
Don’t just publish. Multiply reach.

✅ 7. Upgrade Your Old Content

Old content still ranks but is it optimized for today’s AI crawlers or Bing-powered RAG tools? Add FAQs, commentary, video embeds make your content citation-worthy.

✅ 8. Encourage UGC & Earned Media

Get people talking about your brand. A tweet, a Reddit mention, a Substack shoutout all of these become part of the “digital fabric” that LLMs pull from.

Final Thoughts: SEO Isn’t Dying, But Snake Oil Might Be

The truth? SEO is evolving, not evaporating.

We’re not in a world where it’s “Google vs. AI.”
We’re in a world where those who understand both will rise.

So let’s ditch the fear.
Let’s ditch the hype.

Let’s stick to strategy, tech literacy, brand growth, and content quality.
That’s what’ll get you seen whether it’s by a human or a robot.

Big respect to Daniel Foley Carter for calling out the noise and encouraging deeper thinking in our space.

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *