marketing artificial intelligence
GlobeNewswire
Published on : Sep 4, 2026
AI search visibility may vary significantly depending on which generative AI engine a buyer uses, according to new measurement data from Treyci. The AI visibility intelligence company found that different AI engines can recommend the same brands at substantially different rates, potentially making single-platform visibility checks an unreliable measure of brand performance.
In one measured B2B software category, Treyci found that one AI engine referenced tracked brands in 81% of purchasing-related answers, compared with 43% for another engine. The difference represents nearly a two-fold visibility gap despite the engines receiving comparable buying questions during the same month.
The findings are based on Treyci's measurement methodology, which evaluates buying-intent questions across ChatGPT, Perplexity, Gemini and Grok. The company runs approximately 100 questions per category, including queries such as "best X for mid-size teams," alternatives and pricing comparisons. Each prompt is repeated three times per engine every month, producing more than 1,200 scored AI responses.
The analysis highlights a challenge that traditional search monitoring does not fully capture: AI-generated answers can vary between both platforms and individual sessions.
According to Treyci, the same engine can return different vendor lists when presented with the same buying question at different times. This means a single screenshot or isolated AI search result may demonstrate that a brand appeared once, but does not necessarily establish a reliable visibility trend.
The findings add measurement complexity to the emerging discipline of Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO). Brands increasingly need to understand how they appear when prospective customers ask AI systems for recommendations, alternatives or product comparisons.
Unlike conventional search results, AI-generated responses can synthesize information from multiple sources. Treyci says its analysis found that AI engines frequently cite third-party review platforms, comparison content and industry publications rather than vendor websites.
That dynamic could shift how marketing teams approach AI visibility. Publishing content on a company website may remain important, but third-party sources that AI systems rely on can also influence whether a brand enters a generated recommendation.
Treyci also found a gap between adoption and measurement. In a scan of 100 B2B SaaS companies, 41 had published an llms.txt file for AI crawlers, while relatively few could quantify whether their AI visibility efforts had changed recommendation frequency.
For B2B marketers, the central issue is moving from anecdotal AI visibility checks toward repeatable measurement. Tracking one engine or one query can obscure differences between AI systems and the variability of individual responses.
Treyci's methodology instead treats AI visibility as a distribution that needs to be observed across multiple engines and repeated queries. This approach could become increasingly relevant as AI assistants influence early-stage purchasing research.
The commercial challenge is also different from conventional web analytics. When a buyer sees a company in an AI-generated shortlist but does not click through, the brand may receive no corresponding impression, session or referral record.
As a result, marketing teams may need dedicated visibility metrics to understand whether their brands are being considered before measurable website activity occurs.
Get in touch with our MarTech Experts
Looking to publish a press release, guest article, interview or podcast? Connect with us.
GET FEATURED