GEO, User Reviews and LLMs: Visibility Shifts Ground

GEO, User Reviews and LLMs: Visibility Shifts Ground

LLMs draw on user reviews over SEO

Nothing beats lived experience, even when it is someone else’s. With the rise of large language models and conversational AI, the visibility of travel brands - from destinations and hotels to activities - may increasingly hinge on user reviews and community content. 

That shift is no longer a hunch; the data now back it up. Today’s AI platforms appear to be giving greater weight to signals rooted in user experience to underpin answers and build trust. 

“Business listings make up 48.7% of ChatGPT’s citation mix, a share that dwarfs anything equivalent in a Google SERP,” said Shane Tepper, co‑founder of Resonate Labs, citing a 2025 Yext study.

A model such as OpenAI’s can cite a review directly, yet those first‑hand accounts posted on forums like Reddit are also among the ingredients used to train models. 

“Reddit has signed AI licensing agreements totalling around $203 million in disclosed contracts, anchored by a reported $60 million per year deal with Google,” Tepper said. 

“Sentiment and claims from Reddit threads are baked into model training data and influence what the AI says about a brand even when Reddit isn’t cited at query time. A single forum post can shape the narrative summary without ever appearing as a visible source.”

In this new ecosystem, Reddit has emerged as the most‑cited community domain for answer engines, ahead of YouTube and Wikipedia. With a 3.11% citation rate across platforms, Reddit comfortably outpaces TripAdvisor (0.43%) and LinkedIn (0.41%), according to analysis by Profound in October 2025.

SEO’s blind spots

This phenomenon, observed empirically if not spelled out by the makers of the black‑box LLMs, shows that this new mode of search is far more than turning natural‑language queries into tidy keywords. 

Conversational tools are stepping in where the classic ranked blue links increasingly fall short. Even before ChatGPT’s arrival, it had become common to add “Reddit” to a Google query to bypass, as much as possible, the futility of a first results page where over‑optimised content battles it out with keywords stuffed into h2 and h3 tags, low‑value outbound links and absurdly long articles designed to prompt scrolling in line with a WordPress SEO plug‑in’s diktats.

Should GEO - for everyone’s sake - replace SEO? For now, conversational AI promises to fill in SEO’s blind spots precisely by ignoring the optimisation rules - in truth, proxies for relevance such as keyword density or content length - that search engines have long imposed. 

“User‑generated content tends to be specific because the people writing it are describing actual usage. 

"They name the tier, the error code, the exact feature that broke. Marketing copy tends toward abstraction because that’s what brand teams are trained to produce. 

"The distinction is about which content contains concrete, passage‑extractable details that models weight toward. Authenticity is a story humans tell about the outcome, not the mechanism driving it,” Tepper said.

For practitioners, the fundamentals do not change dramatically, said Aude Naveilhan, Country Manager France at Mirai. The need for visibility means SEO and GEO will rest on common foundations: “Keep to the same practices you use for SEO: structured data and highly granular data.”

The challenge now is to shift from “ranking” to “recommendability”. “Internal tests were run: asking the same question 1,000 times to Claude and to Gemini returned the same brands in only 2% of cases,” Naveilhan said, highlighting the importance of each query’s context - what was asked before, and how previous answers were received - which means there is no longer, in absolute terms, a single “best” result. 

The point is no longer to hide flaws, because AI seeks an honest assessment and draws on Reddit for both positive reviews (5% of citations) and negative ones (6.1%). 

Hoteliers should therefore focus on managing their Google Business Profile listing - a key trust anchor for many LLMs - and encourage an authentic presence in niche communities to become, in AI’s eyes, a genuine source of truth, an experience others can learn from.