When Mitesh Shah talks about the future of AI commerce, he does not start with chatbots or shopping assistants.
He starts with trust.
“A consumer can forgive Netflix for recommending the wrong movie,” Shah says. “They won’t forgive an AI agent for buying the wrong $2,000 plane ticket.”
That distinction, he argues, is why the next phase of artificial intelligence will not simply be about smarter models. It will be about building systems consumers trust enough to let AI act on their behalf financially.
As the technology industry races toward a future where autonomous AI agents can independently shop, compare products, negotiate pricing, coordinate fulfillment, and execute transactions, one problem has quietly emerged as the defining bottleneck for the entire sector: how to make those systems commercially trustworthy at scale.
That challenge has rapidly become one of the most important infrastructure problems in the global digital economy — and Shah has emerged as one of the product leaders working at the center of it.
Over the past decade, Shah has built a career across Amazon, Uber, Block, and PayPal focused on the invisible systems operating beneath digital commerce itself: marketplaces, merchant infrastructure, checkout systems, payments architecture, automation frameworks, and increasingly, AI-native transaction infrastructure.
Today, his work sits at the intersection of AI, commerce, and financial trust systems — an area many industry observers believe could determine how quickly autonomous commerce becomes mainstream reality.
The Hidden Problem Beneath AI Commerce
Much of the public conversation around AI commerce has focused on consumer-facing experiences: conversational shopping, AI assistants, recommendation engines, and autonomous purchasing agents.
But underneath those interfaces lies a much harder systems challenge.
Existing financial infrastructure was designed around explicit human intent. Humans click buttons. Humans approve transactions. Humans authorize purchases.
Autonomous AI agents fundamentally disrupt that model.
Once software agents begin independently evaluating products, selecting merchants, optimizing pricing, coordinating delivery, and initiating transactions on behalf of users, entirely new questions emerge around accountability, fraud prevention, authorization, explainability, and liability.
What happens if an AI agent purchases the wrong item? How should financial systems verify intent? What evidence should exist if a dispute occurs? How do merchants trust AI-originated transactions? And perhaps most importantly: who remains accountable when software starts acting autonomously inside financial systems?
Industry analysts increasingly believe these trust and governance problems — more than AI capability itself — may determine whether autonomous commerce can scale globally.
That is the problem Shah has increasingly focused on solving.
Rather than treating AI-generated transactions as standard payment events, Shah has advocated for infrastructure capable of preserving contextual decision-making throughout the transaction lifecycle. In practical terms, that means maintaining visibility into the instructions consumers gave AI agents, the products and merchants evaluated, the tradeoffs considered by the system, and the reasoning behind final purchasing decisions.
“Financial systems cannot operate as black boxes,” Shah says. “If autonomous systems are participating in commerce, explainability and accountability have to become foundational.”
That thinking is increasingly shaping how technology companies and financial institutions approach the architecture of AI-driven commerce.
A Career Built Around Large-Scale Commerce Systems
Shah’s work in AI commerce reflects more than a decade spent building systems that quietly power large-scale digital transactions.
After graduating from IIT Bombay and Harvard Business School, Shah joined Amazon, where he became a founding product manager for Buy with Prime, Amazon’s initiative to extend Prime checkout, fulfillment, and returns capabilities beyond Amazon.com into the broader merchant ecosystem.
The initiative represented one of Amazon’s most ambitious attempts to export its logistics and trust infrastructure across the internet. Under Shah’s product leadership, the platform scaled to more than 1,000 merchants and surpassed $500 million in gross merchandise volume within its first several years.
At Amazon, Shah also led infrastructure services responsible for checkout promises and routing systems supporting roughly 120 million packages per week, while helping launch machine-learning models capable of predicting fulfillment risk and dynamically adjusting delivery expectations in real time.
Former colleagues say that experience helped shape Shah’s systems-oriented approach to product development.
“At Amazon, you learn very quickly that trust is operational,” says one former peer familiar with Shah’s work. “Consumers experience trust emotionally, but underneath that emotion is infrastructure — delivery reliability, transaction accuracy, fraud prevention, fulfillment promises. Mitesh naturally gravitated toward those foundational systems.”
Shah later moved to Uber, where he led AI-driven automation systems across customer support and post-trip experiences. His work modernizing Uber’s Activity and self-service infrastructure helped automate roughly two million customer interactions per week while improving operational efficiency and customer satisfaction metrics.
One project involved redesigning how users navigated past rides, support requests, and transactional workflows inside the Uber app.
“The interesting thing about automation,” Shah says, “is that users only notice it when it fails. Great infrastructure disappears into the experience.”
That philosophy continued at Block, where Shah focused on modernizing the company’s Data and AI infrastructure stack. There, he worked on systems enabling machine-learning-driven lending decisions, fraud detection, recommendation engines, customer conversations, and automated financial operations at scale.
He also helped spearhead the company’s broader generative AI platform strategy, introducing enterprise capabilities around model hosting, experimentation, evaluation, and AI infrastructure in collaboration with partners including OpenAI, Databricks, and AWS.
Those efforts supported the launch of more than 20 AI-driven products across the organization.
Across each of those roles, Shah consistently worked on infrastructure problems that most consumers never directly see but that fundamentally shape how commerce functions online.
That systems-level background positioned him early for the rise of autonomous AI commerce.
Building the Infrastructure for AI-Native Commerce
At PayPal, Shah’s work has increasingly focused on preparing commerce infrastructure for a world in which AI agents participate directly in shopping and payments.
Among his most significant initiatives has been leading the development of PayPal’s AI-powered Shopping Feed, an effort aimed at transforming the PayPal app from a transactional utility into a personalized commerce and discovery platform serving more than 100 million monthly active users.
He also helped build PayPal’s enterprise AI personalization platform, consolidating fragmented data systems into a real-time intelligence layer spanning more than 400 million consumers and 30 million merchants globally.
That infrastructure now supports a growing range of AI-driven experiences across personalized offer targeting, intelligent wallet experiences, precision marketing, and emerging forms of agentic commerce.
At the same time, Shah has worked on systems designed to help merchants participate more effectively in AI-driven shopping ecosystems by making catalogs, pricing, inventory, fulfillment workflows, and checkout experiences interoperable with conversational AI environments and autonomous shopping agents.
Why Trust May Become the Competitive Advantage
One of the recurring themes in Shah’s work is the belief that the long-term winners in AI commerce may not necessarily be the companies with the most advanced AI models.
They may instead be the companies consumers trust most.
As regulators, financial institutions, merchants, and technology companies begin examining how autonomous systems should operate inside highly sensitive economic environments, infrastructure questions around accountability, transparency, authorization, and consumer safeguards are becoming increasingly central.
That shift is changing how many organizations approach AI commerce strategy.
Rather than treating autonomous commerce solely as a product challenge, Shah approaches it as a systems and governance challenge — one requiring entirely new operational standards for trust.
Industry peers say that perspective has made Shah an increasingly visible voice in conversations around the future architecture of AI-driven commerce.
The Larger Shift Reshaping the Technology Industry
The significance of this work extends far beyond any single company.
The standards now emerging around AI-driven payments, transaction explainability, merchant interoperability, identity systems, and authorization frameworks may eventually shape how autonomous economic activity functions across the global internet.
As AI agents become more capable of independently acting on behalf of consumers, the companies and product leaders helping define those trust systems are likely to exert outsized influence over the next generation of commerce infrastructure.
Shah’s career trajectory reflects a broader transformation happening across the technology industry itself: the convergence of AI, financial systems, digital identity, marketplaces, and commerce infrastructure into a new operational layer for the global economy.
As autonomous commerce moves from experimentation toward mainstream adoption, Shah has emerged as one of the product leaders helping shape how that infrastructure may ultimately work.
