LLMs may notice HOW we communicate, not just keywords

Why semantic patterns, repeated messaging, and content ecosystems matter more than most people realize.

After last week’s LLM Visibility Test results, I wanted to spend a little more time sharing my biggest takeaways. This gets language geekly fast, so if that’s not your thing BUT you know someone who may be interested, feel free to

and come back next week.

And now it’s time to use my graduate education that I’ll never fully pay off. Hmmm.

---

After running an LLM Visibility Test across ChatGPT, Claude, Gemini, Copilot, and Perplexity, one pattern became impossible to ignore: The systems weren’t just recognizing topics. They also appeared to recognize repeated communication patterns across years of public content. That realization may have major implications for GEO, AI visibility, podcast discoverability, and professional reputation online.

LLMs Are Picking Up Patterns, Not Just Keywords

One of the strongest patterns across the experiment was that the LLMs weren’t simply pulling isolated keywords. They appeared to recognize recurring ideas, semantic relationships, and communication patterns across my public work.

Several of the LLMs independently described my work as:

• practical

• human-centered

• anti-hype

• sustainable

• connection-focused

• authenticity-driven

Those descriptions weren’t coming from a single bio page. They appeared to emerge from repeated communication patterns across:

• podcast interviews

• newsletters

• articles

• platform descriptions

• social posts

• transcripts

• long-form conversations

This suggests that AI visibility may increasingly involve more than:

• keywords

• backlinks

• metadata

• SEO structure

Repeated communication style may also shape how LLMs interpret the story around someone’s work online.

---

This Substack is where I experiment with many ways to communicate with the help of LLMs like Claude, ChatGPT, etc. If that’s your jam,

---

Semantic Patterns May Matter More Than Individual Posts

This makes sense when you think about how semantic analysis works across AI systems. These tools are designed to detect broader language relationships and repeated conceptual patterns, not just isolated keywords. That means the systems may gradually build larger thematic associations around someone’s work over time.

For example, many of the themes the LLMs associated with me, like:

• discoverability

• communication clarity

• AI-assisted workflows

• sustainable creator systems

were not coming from a single article or profile page. They were emerging from years of repeated language patterns across podcast episodes, titles, transcripts, newsletters, interviews, and platform descriptions. In other words, the broader content ecosystem around someone’s work may now influence how LLMs understand and categorize expertise online.

Podcast Ecosystems May Now Influence AI Visibility

This experiment also reinforced how much podcast titles, transcripts, metadata, newsletters, and repeated copy, posts, etc appear to shape how LLMs interpret someone’s content online.

That overlap between communication clarity, discoverability, semantic consistency and AI-readable expertise is something I’ve been increasingly exploring through my podcast optimization and visibility work with business-focused podcasts and community-driven brands.

I started noticing these patterns while restructuring podcast titles, descriptions, metadata, and cross-platform messaging for clients. One recent example involved reworking 92 podcast episode titles to make them clearer for both humans and increasingly AI-readable systems.

GEO Is Becoming More Humanish, Not Less

One of the strangest things about modern AI visibility is that it may actually reward clearer, more human communication over disconnected optimization tactics.

The strongest patterns across the LLM experiment were not:

• keyword stuffing

• aggressive SEO language

• polished branding jargon

The systems consistently responded more strongly to:

• repeated conceptual clarity

• authentic communication patterns

• cross-platform consistency

• sustainable thematic repetition

That may be one of the biggest shifts happening right now in GEO and AI visibility. Not less human communication. More semantically consistent human communication. And honestly, that’s probably a healthier direction for the internet overall.

What do you think?

More next week,

Stephanie