A Practical Guide to Being Visible in Both Search and LLMs: Part 1
For decades we all focused on how to get our brand to rank in Google. Today, we’re switching course: now we want our websites, blogs, and articles to be visible in AI-generated answers too. We want to be the source AI applications such as ChatGPT, Perplexity and others pull from when someone asks a question about our product, solution or industry.
So, how do you become a top notch information source for AI?
In our experience, by tackling semantics.
What does semantics have to do with AI search? A lot, it turns out. Semantics is the branch of linguistics and logic concerned with meaning. It acts as the bridge between raw data and LLMs, improving the explainability and coherence of LLM's outputs. LLMs are the engines, the fundamental architecture, behind ChatGPT and the like.
Let's break it down.
What’s an LLM?
Just to make things easy, we can think of a Large Language Model (LLM) as an incredibly well-read librarian, nerd beyond nerd, someone who has spent her entire life studying every book, article, and website ever written (okay, she is superhuman).
When you ask her a question, she doesn't run off to search for the best answer in the library. She already has everything memorized. She can instantly put together the most likely and relevant response to your question by drawing on patterns, context, and connections she has seen across countless texts.
LLMs work the same way. They generate responses by predicting what words or ideas should come next, based on everything they've learned from billions of examples. It's not guessing; it's pattern recognition on a massive scale.
How do LLMs “think”?
To understand how LLMs think and operate, we need to look at the building blocks they work with. These are entities, embeddings, cosine similarity, and monosemanticity.
In this post, we're focusing on entities. In our next posts, we'll tackle the others, one by one.
What’s an entity?
An entity is any specific person, place, thing, or concept. For example: "San Francisco", “climate change”, "Fettuccine Alfredo" or "John F. Kennedy.” Entities are how LLMs understand what your content is really about. They go beyond keywords. They focus on relationships between ideas.
So, let's say you're in the shoewear business and you specialize in hard-to-find sizes and widths. Instead of obsessing over the keyword "best shoes," your content should show that:
- You're talking about why it's important to wear the right size shoes.
- You talk about available sizes, widths, calf circumference (for boots), adjustable straps, the materials the shoes are made from etc.
- You mention your branch addresses, neighborhood names, online store, maybe even reviews.
In short: you're painting a full semantic picture.
Why Context Matters for LLMs
Let's say you mention "jaguar." Are you talking about the animal, the car, or the football team? Context clears that up. LLMs rely on surrounding words and related entities to figure out what you mean.
So when you're writing, give the model clues. Help it build a clearer picture of your topic. Here are 5 ways to do it.
%201.png)
1. Build Topic Authority
Don't just write one blog post about shoes in difficult-to-find sizes and widths. Build a whole hub of content:
- Types of widths, why shoe width matters for foot health
- How nowadays you can find shoes in any size and width and still be classy
- Reviews of different brands: comfort, available sizes and widths etc.
- Why it doesn’t have to be hard to find tall boots in extended sizes with narrow calves
All this tells AI search engines and LLMs: You know your stuff.
2. Link it all together
Internal linking helps connect the dots between related topics. It strengthens the web of entities on your site, making it easier to understand your niche.
3. Use natural, entity-rich language
Instead of saying, "Tall boots are awesome," say:
"Tall leather boots have a long global history, evolving from utilitarian use to a fashion statement, to become a winter staple we all cherish today, in all sizes and widths."
That sentence has clear, specific entities: Tall boots, fashion statement, all sizes. The kind of stuff LLMs love.
4. Format for answers
Structure your content in ways AI tools can easily read:
- Use Q&A format
- Write in bullet points or lists
- Be clear and specific
These formats are exactly what LLMs are trained on.
Help LLMs find you on the map of the internet
When an LLM answers a question, it's pulling from its internal "map" of the internet: a semantic network of entities, ideas, and relationships. If your content is part of that map (because it's semantically rich, structured, and entity-aware), you're more likely to get picked up.
If not? Well, you might end up invisible even if your keywords are spot-on.
TL;DR? Here's your summary:
We're no longer optimizing for just search engines. We're optimizing for LLMs. So that AI search engines start recommending our brand to users. So start thinking in terms of:
- Entities
- Context
- Relationships
- Structured content
In our next blog post, we'll explore embeddings which can help you unlock even more LLM visibility.
Until then, go ahead and be semantically awesome.
Questions? Comments? Write to us at answers@limy.ai.
We’re not as fast as an AI search engine, but we’re working on it.