Your AI assistant just bought something for you.

Not because you asked it to, but because it noticed your partner's birthday approaching, remembered them describing that perfect cup of sencha they had in Kyoto, checked your usual gift budget, and found three tea masters who import the exact varietal from that region. All with proper brewing instructions and handmade ceramics. By the time you read through its recommendation of the spring harvest set, it’s already been ordered. 30 seconds, done. And honestly? It's more thoughtful than another generic tea sampler from the grocery store.

This isn't some far-off example. It's happening now, and it's reshaping commerce in fundamental ways. Agents aren’t just a UX skin or a chatbot added on top of the same old websites. They stand to re-architect the store itself: AI agents that exchange buying and selling intent, preferences, and money with zero human interaction and clicks. What we’re seeing is a categorical break from ecommerce as we know it, and while the storefront might still be there for people, the research and purchase activity will increasingly happen in this new, programmable AI channel.

The shift from ecommerce to agentic commerce

We all understand storefronts. They used to be physical places where you'd walk in, browse products, talk to someone who knew what they were selling, and leave with something you wanted. Commerce was local, personal, and built on relationships between customers and shopkeepers who understood shoppers’ preferences.

Ecommerce took that concept online. Instead of a geographic front door, you had an algorithmic one. Brands published product catalogs, replaced knowledgeable staff with search filters, and optimized relentlessly for rankings and conversion rates. Amazon and Shopify democratized global reach, but they also reduced differentiation. Every brand became just another page in an endless catalog. Every page started looking the same.

Agentic commerce represents something fundamentally different. If ecommerce digitized the storefront, agentic commerce digitizes the buyer, seller, and the decision-making logic between them. Instead of humans navigating static websites, you have intelligent software representing each party and negotiating in real-time to get to a purchase. The buyer doesn't even need to be present for most transactions.

This creates entirely new challenges around trust, data access, and interfaces that can communicate intelligently with each other. For merchants, it's both a rearchitecture of how to reach and serve customers and the next multi-trillion dollar commerce opportunity.

The evolution of commerce

The value exchange of agentic commerce

Agentic commerce operates through a new kind of value exchange between three intelligent entities: consumer agents, merchant agents, and payment infrastructure.

Consumer agents serve as sophisticated shopping assistants that understand context, preferences, and purchasing history. They generate "buyer intents": structured, detailed requests like "Find me a housewarming gift under $150 for my minimalist friend who's passionate about sustainability, delivered to Portland by Thursday."

Merchant agents respond with "seller intents": dynamic offers based on real-time inventory, delivery logistics, pricing constraints, and brand positioning. Instead of static product pages, they craft intelligent responses: "We have three pieces from our sustainable collection, all available for Wednesday delivery to Portland, and we can include a handwritten note about the environmental impact."

Payment infrastructure handles "payment intents": the permissions, spending limits, fraud detection, and execution mechanisms that enable agents to complete transactions autonomously on behalf of people.

Consider this exchange:

Order intent (request): "Find a gift under $100 that ships by Friday and fits my friend's style. Here’s a photo we took together recently."

Seller intent (response): "Here are three items we think she’d love, with gift wrap and overnight shipping included, plus a 10% discount if you buy two (since you’re a returning customer)."

Payment intent (transaction): "Authorize $87.50 for Item B with expedited shipping to arrive Thursday, contingent on gift wrap availability."

This isn't a search result. It's an exchange between agents that results in a structured commitment each party can verify.

Building this system requires new capabilities. Consumer agents need to understand preferences and translate them into actionable buying intents that adapt to timing, occasion, and evolving behavior patterns. Merchant agents need upfront access to structured, machine-readable brand data rather than marketing copy buried in HTML. Payment systems need transparent, verifiable logic and documentation that builds confidence for both users and their agents. Most importantly, all three components need to establish trust in ways that traditional commerce never required.

Trust in the agentic world

Trust doesn't disappear in agentic commerce: it evolves.

Trust in commerce has always been layered: personal relationships, brand reputation, social proof, and institutional guarantees all play roles. What changes in agentic commerce is who evaluates these trust signals. Instead of consumers researching reviews, comparing prices, and weighing brand reputation themselves, their agents will increasingly do this evaluation automatically and at scale.

This shift forces brands to think like software companies, not content companies. Consumer agents will prioritize brands that provide structured APIs over marketing copy, demonstrate pricing accuracy through data feeds over promotional claims, and maintain real-time inventory systems over aspirational product catalogs. The question isn't just whether your brand story resonates, it's whether your systems can reliably serve accurate information to algorithms evaluating hundreds of alternatives in milliseconds. Your software infrastructure becomes part of your brand identity.

Brands that build reliable merchant agents, maintain transparent decision-making logic, and deliver consistent technical experiences will earn this automated trust. Those that continue operating like traditional content publishers risk becoming invisible to the agents making purchase decisions on behalf of millions of consumers.

The fundamentals of trust remain the same. But proving trustworthiness now requires the reliability, transparency, and technical precision that software companies have always needed to survive.

Inside the agentic stack

Consumer Agents

The AI products we use daily (ChatGPT, Claude, Gemini, Perplexity) are evolving beyond question-answering into systems that reason, plan, and take action. The transformation from chatbot to agent happens when you provide AI models with tools: browser access, API keys, persistent memory, and autonomy to work toward broad goals.

Another scenario: Your AI notices you're running low on household essentials. It reviews your purchase history, compares prices across retailers, identifies optimal deals on your preferred brands, and places orders for delivery. Or perhaps it's managing your entire grocery workflow: tracking consumption patterns, learning preferences, and maintaining a fully stocked pantry without your involvement.

As agents become autonomous decision-makers, fundamental questions emerge:

• What personal and financial data can they access?
• Under what circumstances can they act independently?
• How do they choose which merchants to engage with?

These aren't just interface design challenges. They're architectural questions that will determine how value flows through the emerging AI commerce stack.

Users will gravitate toward agents that consistently deliver superior outcomes: faster fulfillment, better prices, stronger alignment with stated preferences. Given that consumers are already paying $20+ monthly for AI subscriptions, they expect these tools to perform high quality, accurate work. In this context, traditional ecommerce tactics like paid advertising, search reranking, and promotional placements start looking like attempts to steer agents toward suboptimal decisions.

Merchant Agents

Most brand and product data today exists as human-targeted content: SEO pages stuffed with keywords, product descriptions crafted by marketing teams, and checkout flows designed around manual form completion. Agents need something different. They don't browse and compare like humans do; instead, they analyze data across multiple sources to find optimal matches.

Current ecommerce infrastructure creates friction for agents that need structured information, not marketing prose. Merchant agents solve this by giving brands a direct way to communicate with consumer agents: sharing product details, pricing rules, and availability in formats that machines can easily process.

To participate effectively in agentic commerce, merchants need systems that:

• Expose real-time catalog data including inventory levels, pricing, and fulfillment options
• Support conversational, semantic search rather than rigid keyword filtering
• Can ask clarifying questions when order intents are ambiguous
• Provide rich visual and contextual information for both agents and users
• Operate at discoverable endpoints (like ai.yourbrand.com) that agents can locate and query

These systems don't replace existing websites: they operate alongside them as parallel interfaces optimized for programmatic interaction. They make businesses addressable through logic rather than just human navigation. This matters because brands want more than traffic. They want agency, control, and a real conversation, just like the shopkeeper who once greeted every customer by name.

When a buyer agent asks "What products would work for this user's specific needs?" a merchant agent should answer with a clear, actionable offer, not a list of links. Price, inventory, fulfilment, and brand story arrive in one structured reply that a consumer agent can trust and act on

The brands that develop these capabilities first will establish ownership in the agentic channel. They'll control how their products are presented to consumer agents, rather than leaving it up to third-party platforms or aggregators. In a world where software speaks to software, the brand that speaks well is the one that wins.

Payment Infrastructure

Once the buyer and seller agents agree, they need a checkout flow built for them, not for people. No carts to fill, no forms to wrangle, no ‘review your order’ loop. The buyer agent applies the right payment method, captures loyalty perks, selects the fastest shipping, and pushes a receipt to the user, all in one atomic call. What used to take several minutes and a dozen clicks will now finish in under a second.

To move money, agents still need rails that are safe, verifiable, and programmable. This requires specialized payment infrastructure designed for autonomous commerce. Four models are emerging to handle agentic payments:

• Human-in-the-loop: Agents handle research and recommendations, but humans complete the actual purchase process.
• Delegated payment methods: Agents receive conditional access to user payment credentials within predefined spending limits and merchant restrictions.
• Agent-owned virtual cards: Users pre-fund virtual payment instruments that agents control, with programmable usage constraints.
• Cryptocurrency wallets: Agents hold digital currency and transact directly using smart contract rules.

Visa and Stripe are both moving fast to enable several versions of this. Visa, one of our launch partners, has built a new suite of Intelligent Commerce APIs that allow for agents to use delegated authentication tokens from user owned credit cards. And Stripe recently demoed a new Order Intents API that provides an abstraction on top of their traditional Payment Intents API. Given a product URL, their system deploys an agent equipped with a virtual credit card that can navigate to the specified website and complete checkout autonomously.

These developments are exciting and signal rapid progress in the payment layer. But they all depend on merchants creating systems designed for interfaces that agents can actually use. The payment infrastructure is evolving faster than merchant systems, and that gap represents both the biggest opportunity and the biggest bottleneck in agentic commerce.

Just as Shopify gave millions of small businesses a way to own their digital aisles, New Gen aims to give merchants a way to own the AI interactions of the future. Control the surface, own the data, maintain the customer relationship. That is the rallying cry for brands in an agentic world.

Merchant agents are the bottleneck

The biggest constraint in agentic commerce isn't payment technology, it's merchant readiness. Payment infrastructure is rapidly evolving to support consumer agents, but most brands lack the structured systems that make automated purchasing possible.

Most product data today remains trapped in human-centric interfaces: disorganized SEO content, fragile product feeds, and legacy content management systems. While adequate for traditional websites, this infrastructure is inadequate for agents that need to handle natural language intent, reason through options, negotiate terms, and make confident purchasing decisions.

Merchant agents transform ecommerce from static websites into dynamic, programmable interfaces. Instead of just displaying product inventory and buy buttons, they expose seller intent. This is the business logic that governs pricing, availability, bundling options, and fulfillment choices. When a consumer agent expresses buying intent like "find a gift under $100 that ships by Friday," the merchant agent can respond with bespoke, contextual offers that reflect both what's available and what the seller wants to promote.

This turns browsing into a conversation. Consumer agents can ask follow-up questions, explore alternatives, and negotiate terms. Merchant agents can offer upgrades, suggest complementary products, or provide discounts based on the specific buying context. The result is commerce that feels more like working with a knowledgeable salesperson than navigating a catalog.

Making this work requires deliberately building standardized frameworks that are:

• Discoverable: Consumer agents can locate merchant agents at predictable endpoints like ai.yourbrand.com
• Structured: They provide machine-readable data and business logic rather than marketing copy that requires parsing
• Reliable: They deliver consistent responses and accurate information that agents can depend on for decision-making

When merchants expose their catalogs, pricing strategies, and fulfillment logic through these intelligent interfaces, they gain direct access to consumer agents. This represents a fundamental shift from the current world of SEO optimization and paid advertising. Instead of competing for search rankings or marketplace placements, brands become active destinations that agents can visit, evaluate, and purchase from directly. High-quality, structured data becomes the new competitive advantage.

The opportunity ahead

The changes happening with AI represent a fundamental shift in how commerce operates. Consumer agents are demonstrating sophisticated capabilities. Payment infrastructure is evolving to support autonomous transactions. The missing piece is merchant-side systems that can engage meaningfully with these emerging agents.

The brands that recognize this gap and act on it won't just gain access to a new sales channel. They'll help establish the foundational standards that govern agentic commerce as it scales. More importantly, they'll evolve from being content businesses to companies that acquire the capabilities of software companies. This will in turn allow them to interface directly with AI systems representing hundreds of millions of consumers.

The current web of static storefronts is giving way to something more powerful: a network of intelligent agents that can reason, negotiate, and transact when mutual value exists. The question isn't whether this future will arrive; it's when brands will be ready to participate in it.

At New Generation, we're building the infrastructure to make this transition possible. If you're ready to position your brand for what comes next, we’d love to talk. Email us. We help global brands build for the AI-native future.