The State of Search Visibility 2025 – How LLMs Are Reshaping Brand Discovery
What We Learned About Brand Visibility from Testing Vans in 7 Major AI Models
In the age of large language models (LLMs), Google is no longer the only algorithm that shapes brand discovery. ChatGPT, Claude, Gemini, and even niche players like Grok and DeepSeek are now influencing what consumers see, trust, and remember — without a single click or SERP involved.
To understand how this shift impacts brands, we ran an experiment using Seshes.ai, our new LLM Brand Tracker platform.
Our test brand?
Vans — a legacy streetwear and skate icon with deep roots in both culture and commerce.
The Setup
We prompted 7 leading LLMs with hundreds of queries simulating real buyer behavior — from casual browsing to product research to culture-driven trends.
We tracked every brand mention, hallucination, citation, rank, and tone.

Key Takeaways
1. LLMs Don’t Echo Google
There was only 43% overlap between top LLM results and traditional SERPs. Brands ranking high in Google often didn’t show up at all in AI-generated answers — and vice versa.
2. Hallucinations Are Real — and Risky
Across 2,310 responses, 23% contained hallucinated outputs. That includes:
- Fictional products and collabs (e.g. “Vans x Thrasher Japan”)
- Made-up brands (“Vision Threads”)
- Confident but false claims about features or availability
3. Prompt Tone Is Everything
Changing just a few words radically changed the LLM output:
- “Iconic skate brands” → Vans, Converse, Nike SB
- “What’s hot in Gen Z culture?” → Palace, Stussy, New Balance
- “Rebellious streetwear brands” → Palace, Vision, Carhartt WIP
4. Vans Dominated — But Not Alone

Vans ranked #1 in most LLMs and SERPs for skate-related queries.
But culturally tuned prompts also elevated lesser-known or SEO-underrated brands like Palace, Stussy, and Lakai.
5. Adidas Got Ghosted

Despite strong Google visibility, Adidas was largely absent from LLM conversations around skateboarding, rebellion, or cultural relevance.
What’s Inside the Full Report?
- LLM Scorecards for Vans across 7 models
- Prompt logs with hallucination flags
- Buyer journey mapping by funnel stage
- Brand vs. Competitor presence breakdown
- Strategic recommendations for LLM SEO
- Appendices with all supporting SQL-backed data
Why This Matters for Marketers
If your brand doesn’t show up in AI-generated answers, it doesn’t exist in the consumer’s decision journey.
LLMs don’t rely on backlinks — they rely on narratives, memory, and prompt resonance.
We’re entering a new era of LLM SEO, and most brands aren’t ready.
Get the Report
The full Vans Visibility Report is now live — including charts, scorecards, prompt logs, and hallucination examples.
👉 [Download the full PDF or view online]
🔍 Or try your own LLM Brand Scorecard at [Seshes.ai]
Want to see how your brand performs inside AI?
Let’s talk. You might be surprised by what the models think you stand for.