Businesses
Mastercard Expands Agentic Commerce Push as AI Agents Move Closer to Shopping and Payments
02 Oct 2026
Mastercard is building new trust, identity and payment infrastructure for a future where AI agents can search, compare, negotiate and transact on behalf of consumers and businesses.
30 September 2026 — Mastercard is accelerating its push into agentic commerce, as artificial intelligence moves beyond recommending products toward independently completing purchases and payments on behalf of users.
The payments company has expanded its Agent Pay programme with new trust and intelligence capabilities designed to help banks and merchants identify AI-driven transactions, understand what users authorised and assess fraud risk before approving payments. The services form part of the Mastercard Agent Pay Trust Framework, which is built around identity, intent, controls, trusted execution and intelligence.
The announcement comes alongside Mastercard’s new report, A Short History of the Future of Shopping and Payments, which explores how AI agents could reshape retail, payments and customer relationships over the rest of the decade. The report draws on Mastercard research and perspectives from four futurists across the US, Europe and Asia.
From shopping assistance to delegated commerce
Today’s AI shopping tools can already help users discover products, compare options and make recommendations. Agentic commerce takes that model further by allowing software to complete parts of the buying process autonomously within boundaries set by the consumer or business.
In practice, an AI agent could eventually search for a product, compare offers, assess delivery options, negotiate on price and initiate payment without requiring the customer to manually complete each stage.
Mastercard’s report goes further, exploring scenarios in which AI agents communicate with other agents, smart devices influence post-purchase decisions and even physical AI systems interact with retail environments.
The company says more than one in 10 online shoppers could routinely use AI agents to shop and pay by 2030, equivalent to more than 300 million people globally, according to predictions included in the report.
For retailers and payment providers, that would represent a significant change in who — or what — sits between a customer and a transaction.
Trust becomes the central business challenge
The technical ability to let AI agents complete purchases may not be the biggest barrier to adoption.
The harder problem is establishing whether merchants, banks and consumers can trust those agents.
A transaction initiated by a person is relatively straightforward to understand. An autonomous agent creates additional questions: Who authorised it? What exactly was it allowed to buy? Was the amount within the user’s limits? Was the agent compromised? And who is accountable if something goes wrong?
Mastercard’s new framework is designed around those questions.
Its five pillars, identity, intent, controls, trusted execution and intelligence, are intended to establish who or what is acting, what the user authorised, which limits apply and whether behavioural or fraud signals indicate that the transaction should proceed.
This could become commercially important because AI agents may eventually operate at machine speed, initiating transactions far faster than traditional fraud systems or manual reviews were designed to handle.
Mastercard adds intelligence to Agent Pay
Mastercard’s latest Agent Pay expansion brings together identity, intent, behavioural and fraud data to provide banks and merchants with greater context around AI-initiated transactions.
The goal is to help financial institutions distinguish legitimate agent activity from suspicious behaviour while avoiding unnecessary friction for genuine users.
The company is effectively adapting existing payment security principles to a new category of participant.
In traditional digital commerce, payment networks have spent decades developing systems for authentication, tokenisation, fraud detection and dispute resolution. Agentic commerce requires those mechanisms to recognise not only the consumer and merchant, but also the AI system acting between them.
For businesses, that could create an entirely new layer of commerce infrastructure.
AI agents could change how companies compete for customers
Agentic commerce may also reshape retail competition.
Today, brands compete for consumer attention through search engines, advertising, social media, marketplaces and recommendation algorithms. If customers increasingly delegate purchasing decisions to AI agents, businesses may also need to compete for the attention of machines.
Product information, pricing, availability, delivery terms and customer-service policies could become increasingly important inputs for AI systems selecting between competing offers.
That could shift the value of traditional marketing.
A visually compelling advertisement may influence a human shopper, but an autonomous purchasing agent may care more about price, specifications, reviews, delivery reliability and the user’s pre-set preferences.
The commercial question may therefore become not simply whether customers know a brand, but whether AI agents consider that brand the most appropriate option.
Mastercard expands its agentic commerce ecosystem
Mastercard is also working with technology companies and startups to build out the surrounding infrastructure.
Its Start Path programme has introduced an Agentic Commerce & Services cohort, bringing together startups developing technologies around AI-led shopping, identity, payments and commerce.
The company has also highlighted partnerships designed to enable agents to make payments on users’ behalf, alongside work with merchants, fintechs and financial institutions to establish standards for agent-led transactions.
The strategy suggests Mastercard expects agentic commerce to develop as an ecosystem rather than a single product.
Payments are only one part of the process. Identity, permission, fraud prevention, data access, customer service and dispute handling will all need to operate together if autonomous transactions are to scale.
Small businesses could also be affected
The shift is not limited to large retailers.
Mastercard has also explored how AI agents could change commerce for smaller businesses, particularly where owners currently spend significant time handling purchasing, administration, payments and customer interactions.
An agent could potentially compare suppliers, reorder stock, manage subscriptions or help negotiate commercial purchases.
For SMEs, that could reduce administrative work, but it could also create new dependencies on AI platforms and payment infrastructure.
Companies may need clearer rules for how much financial authority an agent receives and which transactions should still require human approval.
Agentic commerce could transform commercial payments
Business-to-business transactions could become another major application.
AI agents may eventually be able to manage procurement processes, compare suppliers, monitor contracts and initiate approved payments automatically.
That could reduce friction in commercial payments, but it would also increase the importance of governance.
Companies would need to define spending thresholds, approved counterparties, data-access permissions and escalation procedures before autonomous systems are allowed to move money independently.
In this environment, payment networks are positioning themselves not simply as transaction processors but as providers of trust infrastructure.
Commerce moves from clicking to delegating
The broader change behind agentic commerce is the transition from assisted shopping to delegated shopping.
Consumers have spent decades becoming accustomed to search engines, recommendation systems and personalised advertising helping them decide what to buy.
AI agents could take the next step by turning those recommendations into actions.
That does not mean people will surrender control over every purchase. Mastercard’s report suggests routine and relatively low-risk purchases may be among the earliest areas delegated to agents.
The commercial opportunity lies in reducing the number of steps between intention and transaction.
But the fewer steps a human takes, the more important it becomes to establish clear identity, consent and accountability behind the software acting for them.
For Mastercard, that is where the next competition in digital payments may be developing: not simply making transactions faster, but making autonomous transactions trusted enough to become ordinary.
About Mastercard
Mastercard is a global payments technology company connecting consumers, financial institutions, merchants, governments and businesses across digital and physical commerce.
Its Agent Pay programme focuses on developing payment infrastructure for AI agents capable of conducting transactions on behalf of users and organisations.
Share
Sara Srifi
Sara is a Software Engineering and Business student with a passion for astronomy, cultural studies, and human-centered storytelling. She explores the quiet intersections between science, identity, and imagination, reflecting on how space, art, and society shape the way we understand ourselves and the world around us. Her writing draws on curiosity and lived experience to bridge disciplines and spark dialogue across cultures.
