Key Takeaways
- Wizard, the native AI shopping platform co-founded by Jet.com's Marc Lore, has integrated Mastercard Insight Tokens and Agent Pay through Stripe Shared Payment Tokens to personalize the journey from discovery to checkout
- A direct partnership between the networks (Mastercard, Stripe) and an independent AI shopping surface signals a new pattern, distinct from the network-plus-LLM and network-plus-marketplace alliances seen so far
- For merchants, the practical question is whether their product data and inventory can be discovered by emerging surfaces like Wizard, which will increasingly gate agent-driven traffic
A Two-Layer Stack Built for Personalized Agentic Shopping
Wizard, the native AI shopping platform co-founded by Jet.com's Marc Lore, has launched a partnership with Mastercard and Stripe.
www.pymnts.comOn Thursday, April 30, 2026, Wizard, the native AI shopping platform co-founded by Jet.com's Marc Lore, announced a partnership with Mastercard and Stripe. Building on an existing Stripe-Wizard arrangement, Wizard is integrating Mastercard Agent Pay via Stripe Shared Payment Tokens while also drawing on Mastercard Insight Tokens. The stated goal is a seamless, trusted agent-led experience from discovery through checkout.
What makes the announcement worth a closer look is the way the stack is split into two clear layers. Insight Tokens act as a signal layer, allowing AI agents to apply geography-based spending insights to search and discovery. If a shopper asks for an oven-safe Dutch oven under a certain price, the agent can weigh local price distributions and category preferences before narrowing the candidates. Agent Pay sits one layer below as the payment primitive that lets agents securely initiate and complete the transaction. Stripe's Shared Payment Tokens act as the connective tissue, carrying scoped authorizations such as currency, maximum amount and expiration window between agent and seller.
Wizard CEO Melissa Bridgeford framed the deal in terms of a missing signal layer.
Personalization has always been Wizard's north star, the idea that your AI shopping agent should know you well enough to make recommendations you actually trust. Mastercard's Insight Tokens give us a powerful new signal layer grounded in real-world spending behavior, and with Agent Pay deployed through Stripe, we can take shoppers from personalized discovery to secure checkout in a single, seamless experience.
The split between signal and payment is intentional. Mastercard, as it laid out on its Q1 2026 earnings call, is upgrading its 3.7-billion-card base into agent-initiated rails through Verifiable Intent and agentic tokens. Stripe, just one day earlier at Sessions 2026, packaged ACP, an Agent SDK and SPT into an Agentic Commerce Suite aimed at making merchants and agents interoperable. Wizard becomes one of the first independent shopping surfaces to plug into both at the same time.
Why Now: Marc Lore's Career Arc Explains the Bet
To understand the logic behind Wizard, it helps to read it as the next chapter in a long arc. Lore co-founded Diapers.com (Quidsi), sold it to Amazon for $545 million in 2011, then launched Jet.com in 2014 and sold it to Walmart for $3.3 billion in 2016. He spent the next five years as president and CEO of Walmart U.S. eCommerce before leaving to run Wonder, a virtual food hall and delivery rollup. With Wizard, which emerged from stealth in February 2026, he returns to the discovery side of commerce.
A pattern runs through these chapters. Each venture targets a structural inconvenience that Amazon or marketplace economics fail to resolve, and tries to address it on a separate layer. Diapers.com served a category Amazon underserved at the time; Jet.com proposed dynamic pricing and basket-aware optimization as a counter to Amazon's fixed catalog logic; Wonder bets on aggregating restaurant brands behind one delivery operation. Wizard, in turn, attempts to stand up a third buying surface outside Amazon, Shopify and Walmart, one designed for agent-mediated discovery from the start.
Bridgeford brings the technical and commercial foundation. She founded Stylust, a conversational commerce startup that Wizard acquired and turned into the basis of its team and stack. The product itself is a URL-based shopping search agent that crosses retailer catalogs, customer reviews, editorial sources and social content, returning a tightly curated set of around five candidates per query. AdExchanger reported that the model relies on take-rate and affiliate revenue, with a deliberate refusal to accept sponsored placements. Best Buy is the first retailer with native checkout, and apparel and beauty are next on the roadmap.
That decision to keep the surface independent is precisely what makes it interesting to the networks. Mastercard's prior agentic deals have leaned heavily on partnerships with general-purpose assistants such as OpenAI, Google and Microsoft, where the buying journey starts inside an everyday conversation. Wizard sits in a different category: a destination people open because they want to buy something. For the networks, supporting both lanes is the rational play.
A Fourth Pattern in Agentic Commerce Alliances
Up to now, agentic commerce alliances have clustered into roughly three patterns. First, networks-plus-general-purpose-LLMs: Visa with OpenAI, Mastercard with OpenAI, Google and Microsoft, Stripe with OpenAI. Second, networks-plus-major-retailers-or-platforms: Visa with Walmart, Mastercard with ASEAN commerce players, Stripe with Shopify. Third, networks-plus-standards-bodies, exemplified by Google's AP2 contribution to the FIDO Alliance. The Wizard partnership opens a fourth quadrant: networks plus an independent AI shopping surface.
For this pattern to work, the independent surface needs two strengths at once: the ability to acquire consumers on its own and the ability to assemble high-quality product data. Wizard has invested heavily in both, with a $50 million Series A from NEA and Accel and a steady stream of press in Fortune, AdExchanger and NRF Next, plus retailer-by-retailer API integrations. Bridgeford's own framing, that "distribution doesn't create a great product," is a strategic statement about staying outside any single LLM or retailer ecosystem.
The networks' upside is straightforward. If Wizard or its peers capture even a few percent of consumer buying journeys, Mastercard and Stripe own the payment share of that surface from day one. Insight Tokens give the discovery surface a signal it cannot generate on its own, while Agent Pay over SPT lets it convert without rebuilding payment plumbing. Because the incentives align cleanly, the same blueprint can extend to other independent surfaces such as Perplexity Shopping, OpenAI's Operator-style agents and ChatGPT Shopping.
What Merchants Should Be Watching for the Next Twelve Months
PYMNTS Intelligence research found that 43% of retailers are piloting autonomous AI, and 81% say they trust AI to function autonomously when proper controls are in place. On the consumer side, 45% would let AI agents complete purchases on their behalf, rising to 54% for Gen Z. At the same time, 95% of consumers report at least one concern about agentic commerce, and 50% say they would trust it more if fraud protections were visible. The takeaway is that the market is no longer in a wait-and-see posture, even if mainstream behavior is still early.
For merchants, the practical agenda starts with agent-readiness of product data. Surfaces like Wizard reward structured product attributes, real-time price and inventory, and citations from reviews and editorial sources. The center of gravity is shifting from classical SEO toward AEO (AI Engine Optimization), and the same product feed often needs to satisfy both independent surfaces and general-purpose LLMs.
The second priority is payment primitive selection. Stripe Shared Payment Tokens are positioning to unify Mastercard Agent Pay and Visa Intelligent Commerce behind a single interface, with BNPL options from Klarna and Affirm coming through the same primitive. Whether a merchant supports SPT increasingly determines whether agent-initiated traffic can convert. With Visa framing Intelligent Commerce as a long-term growth driver on its Q1 FY26 call and Mastercard echoing the message a day later, the rail decision can now be made by working backward from the networks' published roadmaps.
The third priority is placement on independent discovery surfaces. Wizard is still at an early stage and effectively absent from the Japanese market, but in the United States it is plausible that a Wizard-class surface ends up touching a meaningful share of mid-funnel demand. Merchants will need to choose between deep one-off integrations like Best Buy's native checkout and protocol-based coverage such as ACP, where Wizard is also an early partner. For most operators, the protocol path lowers the cost of being addressable on multiple surfaces at once.
Closing Thoughts
The Mastercard-Stripe-Wizard partnership adds a fourth pattern to the agentic commerce playbook: networks paired with an independent AI shopping surface. The two-layer stack of Insight Tokens for signals and Agent Pay over SPT for payments is reusable across other surfaces, which is what makes the announcement more than a single-deal story.
Marc Lore's career arc, from Diapers.com to Jet.com to Walmart to Wonder to Wizard, is a long-running attempt to build new buying surfaces outside the dominant marketplaces. Read in that light, this week's deal is less about one startup and more about the kind of alliance that will define the next phase of agentic commerce. Merchants who treat agent-readiness, payment-primitive choice and surface placement as a single program will be in a stronger position when the next independent surface follows the same template.
