DaVinci Commerce launches Brand Agent and Product Context…

The San Mateo company's expanded platform, available October 1, publishes machine-readable product knowledge to ChatGPT, Meta Muse, and OpenAI's dots — a new discovery surface small B2C sellers can't ignore.

DaVinci Commerce, the San Mateo company founded by former Informatica co-founder Diaz Nesamoney, on Monday expanded its Agentic Commerce Experience Platform with two components, Brand Agent and the patent-pending Product Context Memory, which went live October 1. The pitch is narrow and the timing is pointed: publish enriched, machine-readable product knowledge directly into the agents consumers now actually shop through.

Those agents are no longer hypothetical. Meta Muse’s recent climb to #1 on the US App Store and Instinct’s $1B raise at a $10B valuation have made clear that personal agents, alongside ChatGPT, Gemini, Claude, OpenAI’s dots, and retailer-owned agents like Walmart’s Sparky and Amazon’s Alexa for Shopping, are the new first-contact layer for discovery.

Nesamoney’s framing of the gap is precise. “most product catalogs still speak in technical specifications,” he said, while shoppers now ask agents things like “What foundation won’t turn orange or look cakey after a few hours?” or whether a foundation is “better suited” for sensitive skin. Brand Agent pushes product data through ACP, UCP, MCP services, JSON-LD, and existing PIM feeds. Product Context Memory organizes catalog data alongside ratings, reviews, social conversations, and lifestyle content into a per-SKU knowledge graph with semantic indexing.

A companion component, Product Discovery Intelligence, tracks whether products get retrieved, mentioned, and linked back to the brand’s own page across AI systems, per Forbes.

Two cautions. Runtime Wire notes the launch includes no performance results. And Forbes points out that each AI platform decides independently how much weight brand-supplied context carries against independent reviews, which means publishing structured product knowledge is necessary but not sufficient. Small B2C sellers evaluating this category should treat retrieval and mention metrics as the deliverable, not the input.

The structural read is simpler than the acronyms suggest. Catalogs written for keyword search are becoming illegible to the layer that now mediates consumer intent, and the companies racing to re-encode them are selling translation into a channel whose weighting rules the platforms haven’t finished writing.

DaVinci Commerce launches Brand Agent and Product Context Memory for AI shopping

Sources