AI Models Left a Brand Out of Their Answers 36% of the Time. What That Says About How You Describe Yourself

Marketers have spent the last 18 months coining acronyms for AI discovery, but the more useful question is what actually decides whether an AI model recommends you. A recent report from Language Engine Optimisation advisory LEOPRD analysed over 200,000 AI-generated responses spanning 27 brands across eight major AI models in Australia and the UK. For unbranded category questions, the monitored brand was missing 36% of the time, mentioned but overlooked in favor of someone else 35% of the time, and the primary recommendation just 29% of the time. Answers where a brand appeared but lost to a competitor drew on an average of 9.2 citations, and the report's central finding is that what those sources actually said about the brand mattered more than how often they mentioned it. It also undercuts a common assumption: AI visibility cannot be solved entirely by fixing the corporate website, because the models weigh what third parties say about you, not only what you say about yourself.

For brands and startups, the takeaway is that clear, consistent, distinctive language about who you are and what you do now has to exist well beyond your own pages, in coverage, reviews, partner pages, and the content you publish. That makes brand work more strategic than a logo and a color palette. A sharp positioning statement, a consistent voice, and a steady stream of distinctive content, such as campaign films, illustrated explainers, and recognizable visual assets, give outside sources something specific and repeatable to say about you. A vague or generic brand is easy for a model to blur into its competitors, while a clearly defined one gives it something to quote. This is evidence from a single report across two markets, so treat the percentages as indicative rather than universal, but the direction is clear: be specific, be consistent, and make it easy for others to describe you accurately.