LLM-powered shopping inside the world’s most powerful retail platform is no longer a future concept.

It is already here in the form of Amazon Rufus.

Rufus has been rolling out quietly across Amazon desktop and mobile experiences in multiple markets and, over recent months, the quality of the experience has improved significantly. What looked like an experimental assistant is now becoming a serious layer in how shoppers browse, compare and ultimately buy on Amazon.

This is a commercial imperative you can’t ignore. 

Recent reporting on Amazon’s Q4 2025 performance highlighted the growing impact of Rufus touchpoints, with claims that Rufus influenced billions in revenue during the period and across 2025. Whether you view that as early momentum or a major inflection point, the message is clear - Rufus is no longer a fringe side project. It is becoming a meaningful part of Amazon’s shopping journey and could become one of the main ways people discover and purchase products on the platform. (ir.aboutamazon.com)

For brands, that creates a new challenge and of course a corresponding opportunity.

Until now, there has been no reliable way to understand how your brand is actually performing inside Rufus. Teams could optimise listings and content using best practices, but they could not properly measure visibility, compare competitive positioning, or identify where Rufus was misunderstanding the brand.

That is now changing.

Share of Model now includes Rufus surfaces within its category-level analysis, giving brands a way to review performance not only across leading LLMs such as ChatGPT and Gemini but also within Amazon’s own AI shopping experience. This is a huge step forward for AI commerce measurement.

It means brands can start to assess Rufus visibility with the same kind of framework they already use for broader AI discovery: share of voice, mention rate and average position. As these metrics are viewed at category level, teams can now benchmark against competitors rather than interpreting performance in isolation.

This matters because Rufus does not behave exactly like the classic Amazon search bar.

Traditional Amazon optimisation logic still counts and is super important. Titles, bullets, imagery, reviews, availability and most crucially sales velocity remain critical inputs. But AI-led recommendation experiences introduce another layer. Rufus is interpreting shopper intent in natural language and then deciding which products and brands best match that intent. That changes how we need to optimise for visibility. 

In practical terms, a brand can be strong in classic search and underperform in Rufus if its product truths are not being expressed clearly enough for AI interpretation. This is where Share of Model becomes incredibly useful.

By showing how Rufus sees your brand versus competitors, the tool helps teams identify strengths, weaknesses and therefore missed recommendation opportunities. From there, brands can make targeted updates to content and assets so that Rufus better understands what the product is, who it is for and which shopper queries it should be recommended against.

That includes improving how key brand truths are expressed across the listing and supporting content, tightening the language used to describe functional benefits and making sure the product is clearly aligned to the kinds of real questions shoppers are asking Rufus.

The result is not just better reporting. It is a practical way to improve visibility in one of the most important emerging AI commerce surfaces.

As Rufus becomes more embedded in the shopping journey, the brands that win will not simply be the ones with the biggest budgets. They will be the brands that understand how Amazon’s AI sees them, where that understanding is weak and how to fix it faster than competitors.

That is what makes Rufus measurement so important now and why adding Rufus surfaces into Share of Model is such a significant development for brands serious about AI-led commerce.

If you want to win agentically, start tracking how you surface in Rufus today. 

Author - Stephen Honight, founder of The Lmo7 Agency.