![](\"https:\/\/mms.businesswire.com\/media\/20260604759368\/en\/2823758\/4\/azoma-black.jpg\")
<\/a><\/a>\nRecognised by global retail giants<\/b><\/p>
\nAlready adopted by Mars, Unilever, Beiersdorf, L\u2019Or\u00e9al, and Reckitt<\/b>, the framework defines five pillars critical to AI-driven retail: Completeness, Context, Citations, Correctness, and Customer Acquisition.<\/b><\/p>
\n\"The 5 C's have been instrumental in navigating the AI era,\" said Erica Meagher, Head of Ecommerce at Beiersdorf Canada<\/b>. \"They gave our team a clear, practical framework at a moment when the rules of ecommerce are being rewritten in real time.\"<\/p>
\nAgentic commerce is platform-specific<\/b><\/p>
\nThe framework was developed based on Azoma\u2019s proprietary citation data which analysed tens of millions of AI responses in Q2 of 2026. The findings shatter the myth of a \"one-size-fits-all\" AI strategy, revealing how different AI agents lean on vastly different sources to recommend products:<\/p>
- \n
- \nEarned media dominates voice agents:<\/b> Earned media drives 86.5%<\/b> of citations for Amazon\u2019s Alexa for Shopping and 76%<\/b> for Walmart\u2019s Sparky.<\/li>\n
- \nLLMs lean on retailers:<\/b> OpenAI\u2019s ChatGPT and Google\u2019s Gemini rely far more heavily on direct retailer sources (37.1%<\/b> and 41.6%<\/b> respectively).<\/li>\n
- \nThe category flip:<\/b> The source mix shifts drastically by aisle. Earned media drives 67.6%<\/b> of citations in the Wellness category, while retailer sources dictate 50.5%<\/b> of answers in Food.<\/li>\n<\/ul>
\nThe hidden scale of the data problem<\/b><\/p>
\nAzoma\u2019s citation data exposes a massive backend hurdle: a single product listing on a global retailer can now require 600+ unique attribute values<\/b> to be deemed \"complete\" by an AI. Because AI agents rely on backend metadata rather than consumer-facing marketing copy to determine product eligibility, brands missing these hidden data points are being locked out of AI recommendations entirely.<\/p>
\nDue to AI agents predicting rather than retrieving information, they frequently hallucinate product specs - even when a brand\u2019s official content is 100% accurate. Azoma warns that the only way to catch and correct these fabrications is through systematic, proactive platform testing.<\/p>
\nThe full report, \"The 5 C's of Agentic Commerce,\"<\/i> is available now through Azoma and the Digital Shelf Institute.<\/p>
\nAbout Azoma<\/b><\/p>
\nAzoma\u2019s ambition is to define the merchant-side standard for agentic commerce. Today, Azoma works with many of the world\u2019s largest consumer brands, including Colgate, Mars, P&G, Unilever, L\u2019Or\u00e9al, Beiersdorf, and others. For these enterprises, agentic commerce is not a point-solution problem. It represents a structural shift in how product catalogues are interpreted, governed, and measured as agentic interfaces spread across platforms. What they need is not another tool to manage a single channel, but a shared standard that preserves consistency, compliance, and predictability as this new layer of commerce takes shape.<\/p>
- \nLLMs lean on retailers:<\/b> OpenAI\u2019s ChatGPT and Google\u2019s Gemini rely far more heavily on direct retailer sources (37.1%<\/b> and 41.6%<\/b> respectively).<\/li>\n