CogitX's Sovereign AI, Live at 1 Top Agency Holdco, Gets Retail President

For AI teams evaluating the next wave of enterprise deployment, CogitX AI's retail-focused leadership hire sharpens a growing architectural argument: sovereign models that live inside an organization's own environment versus rented API access billed by the token. CogitX claims its platform is already in production at one of the world's largest agency holding companies, though that account is not named. Ellgass's move from Circana—where he built a global retail media business—signals CogitX intends to take that ownership model to retailers whose AI workloads now span marketing, supply chain, and merchandising.

CogitX AI opened a new chapter in its retail strategy on September 18, 2026, announcing Michael Ellgass as President, Retail. In the role, Ellgass will lead go-to-market across the company’s retail offerings—spanning marketing, supply chain, merchandising, and, with particular emphasis, commerce media. The appointment, distributed through BusinessWire India from Bangalore and carried as a synced brand post, frames the move around a contrarian economic argument: retailers that own their AI outright, rather than rent it by the token, can keep costs flat as adoption grows and build capability that compounds over time.

“

In that role, he worked with retailers including Sam’s Club, Walgreens, Instacart, and Coles.

Ellgass arrives with a career that maps closely to the rise of retail media. The company says he joins from Circana, where as EVP of Global Commerce Media he built and scaled a retail media business from scratch across North America, EMEA, and APAC. In that role, he worked with retailers including Sam’s Club, Walgreens, Instacart, and Coles. Before Circana, Ellgass led category marketing at Walmart—a discipline the announcement describes as having evolved into retail media—building marketing plans with more than 200 brands. He began his career in brand marketing at Kraft Foods. That trajectory gives CogitX not just an executive but a distribution map across the buyers, platforms, and retail brands most likely to adopt commerce-media AI.

The strategic thesis embedded in the release is a direct challenge to usage-based AI pricing. CogitX says the industry currently sells retailers access to intelligence, while it sells ownership of it. The stated operational benefit is that a model living inside a retailer’s own environment becomes smarter with every interaction without a bill that scales with success. That pitch is aimed squarely at finance and technology leaders inside retail organizations that have seen inference costs rise as AI usage expands. If the claim holds, it would shift AI from a variable operating expense to a capitalized capability—an important distinction for retailers with thin margins and high transaction volumes.

CogitX also says the model is already in production at one of the world’s largest agency holding companies, where it is rapidly transforming engagements and workflows with AI powered by the platform. That account is unnamed in the release, and because both source articles are press-release or sponsored-brand distributions rather than independent reporting, the production claim should be treated as the company’s assertion, not independently verified fact. The absence of a named customer and concrete performance metrics leaves the scale and maturity of the deployment open to question, but the existence of even one large agency holdco pilot gives the company a reference point for enterprise sales.

For the retail sector, the appointment lands at a moment when commerce media has become one of the industry’s most profitable growth lines. Retail media networks built by Walmart, Instacart, Walgreens, and others have turned first-party data and digital shelf space into high-margin advertising revenue. Ellgass’s experience building Circana’s global commerce media unit suggests CogitX wants to position its sovereign AI as the intelligence layer underneath those networks—owning the models that power targeting, merchandising decisions, and supply chain optimization. The move could resonate with retailers that are wary of handing data and margin to external ad-tech and cloud providers, but it also requires retailers to invest in internal AI capability rather than simply buying off-the-shelf tools.