What Azoma actually does
Azoma is a three-year-old startup, founded in 2022 by former Amazon executive Max Sinclair and machine learning researcher Timur Luguev, that helps consumer brands optimize and measure their presence across AI shopping assistants. Rather than building another shopping agent to compete with Alexa or ChatGPT, Azoma sits on the brand side of the equation, helping companies like L'Oreal, Unilever, Mars, Beiersdorf, and Reckitt understand how those third-party agents are already influencing which products get recommended and purchased.
That positioning is the point. As retailers and platforms race to launch their own AI shopping assistants, Rufus at Amazon, Sparky at Walmart, Dots inside ChatGPT, Meta's Muse, brands selling through those channels have almost no visibility into how recommendations are generated or how to improve their odds of being surfaced. Azoma's pitch is that someone needs to build the measurement and optimization layer sitting between brands and every one of these agents, rather than brands negotiating that visibility separately with each platform.
The funding and who is backing it
The 11.2 million dollar Series A was led by Puma Growth Partners, with dunnhumby, Twin Path, MaRS IAF, Rank Ventures, and Ignite Ventures participating alongside angel investors from Google DeepMind and Legora. It follows a 4 million dollar seed round announced in December 2025, putting Azoma's total raised at roughly 15 million dollars in under a year of fundraising, a pace that signals investors see the agentic commerce measurement category as moving fast enough to justify quick follow-on capital.
Henri Songeur, investment director at Puma Growth Partners, said Azoma is building a defensible measurement solution in a category it helped create, and pointed to its existing enterprise customer base as evidence the product already works commercially rather than remaining a concept. dunnhumby's participation is notable on its own: as a retail data and media company serving grocers and CPG brands directly, its involvement suggests Azoma's roadmap includes plugging retailer loyalty and transaction data into its attribution models, not just platform-level signals.
The numbers Sinclair is using to sell the category
Sinclair's public pitch leans on two statistics worth treating with appropriate caution. He cites that 350 million people used Alexa for shopping last year, a scale figure for the addressable market Azoma is selling into. He also claims that Instinct, a personal AI assistant that launched in August 2026, already accounts for 1 billion dollars in transactions, a figure reported by the company itself rather than independently audited, and one that should be treated as a sales claim until a third party verifies it.
Sinclair also cited a Walmart data point, that customers using Sparky, Walmart's AI shopping assistant, spend 40 percent more than customers who do not use it. That statistic originates with Walmart rather than Azoma's own measurement, and it illustrates exactly the kind of lift number that brands now need help capturing credit for. If an AI assistant is driving 40 percent more spend, the CPG brand whose products get surfaced inside that assistant has a direct financial interest in knowing which SKUs are winning that recommendation and why.
Why attribution, not the agents, is the investable layer
The strategic logic behind this round is that the shopping agent layer, Rufus, Sparky, ChatGPT, Alexa, Meta Muse, is being built and funded by platforms with far more capital than any startup could match. Competing there directly would be a losing bet. Measurement and optimization sitting across all of those agents simultaneously is a different business, closer to how search engine optimization and attribution firms built durable businesses without ever competing with Google itself.
That is the bet Puma Growth Partners and dunnhumby are making with this round, and it is a reasonable one. Every platform launching its own shopping assistant has an incentive to keep its recommendation logic opaque, which creates a persistent information gap that brands will keep paying to close regardless of which specific agent ends up winning the most market share. Azoma's value proposition survives even if today's leading platforms are not tomorrow's leading platforms.
What this means for brands without an answer here yet
If your organization sells through retail partners and has no current way to measure how AI shopping assistants are influencing which of your SKUs get recommended, that is now a visible gap rather than an acceptable unknown. Azoma's customer list, five major CPG names across beauty, household goods, and confectionery, is evidence that sophisticated brand organizations have already concluded this measurement is worth paying for rather than waiting for platforms to volunteer the data on their own timeline.
The practical next step does not require signing with Azoma specifically. It starts with scoping what internal or vendor-provided attribution you currently have across AI shopping channels, and being honest about whether that scope covers the agents your actual customers are using today versus the ones your team happens to have integrations with already. The 350 million Alexa shoppers and the growth trajectory behind Sparky and ChatGPT's commerce features suggest this gap only gets more expensive to close the longer it stays open, and the brands moving on it now are buying themselves a head start that will be harder to replicate once every competitor has the same measurement in place.
