Commerce Everywhere, Agents in Charge

By Mike Evans, Robert Derow, Raakhi Agrawal, Lachelle Trout, Olof Darpö, and Martin Barthel
Blog Post

How AI Personal Shoppers Will Navigate a World of Embedded Storefronts

A quiet but far-reaching revolution is underway in commerce. Buying is no longer confined to websites, apps, or stores. It is escaping channels and embedding itself across everyday life.

A product appears in a social video.
A recipe is streamed on TV.
A car prompts a driver to pay, recharge, or reorder.

Shopping no longer happens in one place, or even at a single moment. It unfolds across feeds, media, devices, and environments as consumers move through their day. Commerce is being woven directly into the surfaces people already use, turning ordinary moments into moments of decision.

As the surface area expands, managing dozens of touchpoints, offers, and decisions becomes impractical for consumers. Instead, they’re increasingly handing off the work to AI agents . Consumers articulate intent, travel lighter, restock smarter, save time, and expect an agent to figure out the rest.

This marks a turning point for commerce leaders. What is changing is not just where transactions happen, but how buying decisions are made, and who does the work.

What’s Driving the Shift

This shift didn’t happen overnight, and it isn’t the result of a single technology or trend. It’s being driven by deeper structural changes in how commerce operates, and in how consumers navigate buying decisions.

At the center are two forces:

  1. Commerce is no longer confined to channels. It is embedded across everyday surfaces, social feeds, shoppable media, immersive environments, vehicles, and connected devices, so that shopping becomes a capability available everywhere, not a destination consumers visit.
  2. AI personal shoppers are emerging as the organizing layer for this fragmented world. These agents translate intent into action: they research, compare, optimize, assemble baskets, and increasingly execute purchases across merchants and contexts on the consumer’s behalf.

Individually, each force is powerful. Together, they fundamentally rewire how buying decisions are made—and what it takes for brands to win.

The scale of the change is significant. The global e-commerce market is projected to grow from roughly $25 trillion in 2024 to more than $80 trillion by 2030 as commerce expands into new categories and embedded experiences, while AI-driven shopping moves rapidly from experimentation to habit.

The implication is clear: commerce is everywhere, and AI is quickly becoming the way in.

For leaders, the strategic question is no longer, “Which channel are we optimizing?” It is, “How will our brand succeed when buying decisions are initiated—and optimized—by AI agents across every surface?”

What follows examines how these forces show up in the buying journey, and how they converge across five arenas connected by an emerging layer of infrastructure: the universal cart.

The New Mechanics of Commerce

These two forces are changing the mechanics of commerce, altering where buying happens and how decisions move from intent to execution. Here’s how they play out in practice:

Shift 1: Commerce moves from channels to surfaces

What began as a set of discrete destinations has become a capability embedded across the surfaces consumers already use.

Social feeds double as storefronts. Media experiences collapse discovery and transaction. Immersive environments resolve fit and context before purchase. Cars, homes, and connected devices introduce buying moments in motion or in the background.

Five arenas illustrate this shift most clearly:

Across all five arenas, the pattern is consistent: commerce is spreading across surfaces, increasing the number of buying moments while making those journeys harder for consumers to manage on their own.

Shift 2: AI personal shoppers become the front door

As commerce expands across surfaces, consumers are changing how they engage with it.

Instead of navigating dozens of apps, sites, and touchpoints, many now start with an AI assistant. The consumer provides intent, find me a quiet dishwasher, plan a long weekend, keep my household stocked at the best value—and the agent does the work.

These AI personal shoppers research options, compare prices, monitor availability, apply promotions and loyalty benefits, curate baskets, and increasingly execute checkout. What once required a multi-step journey across multiple destinations is compressed into a single act of delegation.

BCG’s research points to this being a “seismic shift” in how commerce is conducted. Shoppers arriving via AI agents tend to have higher intent and clearer constraints, changing engagement and conversion dynamics. The consumer moves from being the operator of every step to the author of the brief.

The Architecture Behind Commerce Everywhere

What has emerged from these shifts is a re-architected commerce model. Buying moments are now distributed across feeds, media, environments, and devices. AI agents act as the connective layer, pulling signals together and turning them into action. Inspiration and transaction no longer need to happen on the same surface.

The New Architecture of Commerce

In this model:

The result is not more complexity for consumers, but less. AI agents make a fragmented commerce landscape usable, while shared cart infrastructure makes it executable.

How This Shift Plays Out Across Five Arenas

The implications of this new architecture become clearer when viewed through the five specific arenas where commerce is already embedded.

How Commerce Plays Out Across Key Arenas

  1. Social & livestream commerce
    If AI agents are becoming the front door, social platforms remain where much of the desire is first sparked.

    Social commerce is already large and growing fast. Embedded shopping on platforms like TikTok and Instagram has turned content into storefronts, with social commerce projected to exceed $1 trillion in the near term and livestream formats expanding rapidly beyond Asia.

    Despite this growth, the buying experience remains clunky. Users tap through links, land on mismatched product pages, and re-enter payment details. AI personal shoppers simplify this journey.

    A consumer sees a creator recommend a jacket in a TikTok video. Instead of hunting through links, they tell their agent: “Find that jacket or something similar for under $200 that can arrive by Friday.”

    The agent identifies the product, checks availability and pricing across social shops, marketplaces, and brand sites. It factors in the user’s size, preferences, and return history, then recommends—or purchases—the best option based on the level of autonomy it has been given.

    In this arena, social content becomes a high-signal source of intent. AI agents validate what consumers see, layering in price comparison, competitive alternatives, ratings, and deeper product information—before buying where value and fit are strongest.

    For leaders, the question shifts from “How do we grow social storefront GMV?” to “When our products appear in social contexts, how easily can AI agents identify, evaluate, and assemble them across merchants?”

  2. Shoppable & commerce media
    Retail media networks and shoppable ad formats are turning media into a transactional channel. What was once built to drive awareness or clicks is increasingly designed to close the sale.

    Often described as the third wave of digital advertising after search and social, commerce media brings media and transaction together, allowing consumers to buy directly from the content they consume. Retail and commerce media are already expected to exceed $100 billion globally and are growing faster than digital advertising, reflecting a broader shift in how attention converts into purchase.

    What has changed is where checkout happens. Platforms now embed transactions directly into media experiences: across ads, content, streaming TV, livestreams, and messaging—so discovery can turn into purchase without breaking context.

    Add AI personal shoppers, and this model becomes more practical. For example, a cooking show might include a shoppable overlay. The viewer’s agent can then parse the recipe, check what’s already in the pantry, build a list of missing items, compare pricing across retailers, and schedule delivery. All the user has to say is: “Yes, do it.”

    In this arena, the “ad” is no longer a suggestion, it is a transactional trigger that an AI agent can act on. Shoppable media turns every impression into a potential checkout, while AI personal shoppers make that scalable, especially when connected to a universal cart and checkout infrastructure.

    For brands and retailers, this compresses what was once a messy, multi-step funnel into a single, measurable touchpoint. Media, data, and offers become one design problem: assets must be readable by machines, offers must appeal to an optimizer, and performance must capture when an AI agent, not just a human, completes the journey.

  3. Immersive & AR experiences
    In categories where fit, aesthetics, and context drive decisions, fashion, beauty, furniture, home, immersive experiences are becoming the new standard. Tools like virtual try-on and “view in my room” allow shoppers to answer critical questions about size, look, and fit before they buy.

    Adoption is accelerating as these experiences move from novelty to expectation. Google’s virtual try-on capabilities offer a clear signal of where this is heading: shoppers can now use full-body photos to see how apparel and footwear might look on them directly within Google Shopping. Similar experiences are increasingly common across cosmetics, accessories, and home furnishings.

    AI personal shoppers turn this visual clarity into action. A shopper tries on a coat, likes the silhouette, but hesitates on price. Instead of restarting the search, they tell their agent: “Find three similar coats that are at least 20% cheaper, warm enough for New York in January, and available before my trip.”

    The agent applies what it has learned, style preferences, fit feedback, climate needs, and past behavior, to search across retailers and return an optimized shortlist, ready for one-tap approval.

    In this arena, the immersive layer becomes where uncertainty gets resolved, and the AI agent becomes how consumers act on that clarity.

    For commerce leaders, the priority is not to chase every new device or experience. It is to invest in immersive moments that materially influence decisions—and to ensure those are legible to AI agents through consistent product identifiers, attributes, and structured signals that enable action.

  4. In-car & IoT commerce
    As cars, homes, and devices become more intelligent, a growing share of commerce happens while people are doing something else.

    Automakers such as BMW and Mercedes-Benz now support in-vehicle payments for fuel, parking, and charging directly from the dashboard. Smart home platforms and connected devices already manage routine replenishment of consumables in the background.

    AI personal shoppers coordinate these moments behind the scenes. They decide when to act, which option to choose, and how to bundle tasks across contexts, often without requiring the consumer’s active input. As Joel Milne, CEO of AutoUnify , notes, “This shift puts the onus on corporations to make their data agent-ready: live, structured, and ingestible, so agents can interpret it reliably and complete transactions without hallucinations or misinformation.”

    In this arena, commerce shifts from interaction to delegation. Consumers are no longer actively shopping, they are authorizing agents to act on their behalf, often invisibly.

    For brands and service providers, success depends on designing offers that work in motion, location-aware, time-sensitive, and easy for agents to combine with other errands while integrating cleanly into automotive and IoT ecosystems. Equally important is trust: when agents act automatically, consumers must feel they are delegating decisions, not surrendering control.

  5. Agentic commerce
    If 2025 marked the rise of the business agent inside the enterprise, 2026 is shaping up to be the year of the personal agent. All major consumer AI platforms are quickly releasing standards for how businesses interact with consumers through AI, accelerating the shift toward agentic commerce.

    In this model, AI assistants do more than just recommend products. They research options, compare prices, apply preferences and constraints, assemble baskets, and increasingly complete purchases on behalf of consumers.

    Consumers express intent directly to an agent, find the best carry-on for frequent travel, restock household essentials, book a weekend trip—and the agent handles the work across merchants and marketplaces, without forcing the user to navigate sites or apps.

    This is already taking shape in familiar environments. Platforms such as ChatGPT, Google Gemini, and Perplexity are embedding shopping capabilities directly into conversational experiences, allowing users to research, evaluate, and move toward purchase without leaving the interface. As these agents connect to universal carts and payment credentials, these environments become places where transactions can close, not just begin.

    In this arena, commerce becomes optimization-led. AI agents continuously weigh price, availability, delivery speed, and reliability to determine what best satisfies a user’s intent at that moment.

    Every brand must first start with strategy to determine if it wants to play a destination game (i.e. drive customers to owned channels to protect margins and brand) or an evaluation game (i.e. get recommended by agents to partake in the growing channel and attract new customers). Most brands will choose the evaluation game—winning in agentic commerce requires being discoverable to machines, not just compelling to humans. Product data, pricing, availability, and fulfillment terms increasingly determine whether an agent selects or passes on an offer.

Universal Carts: The Infrastructure Agents Rely On

If AI agents are the front door and new surfaces are the rooms, the universal cart is the plumbing that makes the whole system work. It’s the infrastructure layer that allows agents to add, hold, and complete purchases across merchants and contexts on a consumer’s behalf.

Early implementations are already emerging. Shopify’s Commerce for Agents toolkit , launched in 2025, enables AI assistants to search across large product catalogs, assemble baskets spanning multiple sellers, and complete transactions directly within conversational or voice experiences, without routing users to individual sites. Platforms like Firmly.ai extend this even further, supporting multi-merchant, multi-cart checkout across platforms and enabling native shopping on any surface.

Here's how universal carts change how shopping works:

Whoever operates this layer, or something functionally equivalent, gains a rare, mission-level view of commerce: how baskets form across retailers, where substitutions occur, and how consumers trade off speed, price, and convenience. That visibility creates a significant advantage for personalization and product strategy, but also raises important questions around trust, transparency, and data governance.

For brands, the strategic question is not whether universal carts will exist. They already do. The real decision is how to participate in ways that protect visibility, margins, and access to data as more of the shopping journey moves through agents and shared infrastructure.

Preparing for an Agent-Mediated Market

“Commerce everywhere” is quickly moving from concept to operating reality. Buying is becoming embedded across feeds, media, immersive experiences, vehicles, and devices. At the same time, AI personal shoppers are increasingly mediating how consumers research, compare, and purchase. Universal carts and shared infrastructure are beginning to stitch these journeys together across merchants and contexts.

Over the next 24 to 36 months, leaders must focus on three priorities as buying decisions move upstream into AI-mediated systems:

  1. Clarify your agentic strategy. Decide where to build proprietary agent experiences, where to rely on horizontal platforms, and how your products will consistently show up in agent-mediated journeys.
  2. Make your business agent-ready. Invest in structured product data, clear policies, and integrations with key surfaces and universal cart ecosystems. Agents will bypass offers that are ambiguous, incomplete, or hard to execute.
  3. Measure missions, not channels. Shift from channel-centric KPIs to mission-centric ones. Measure how often your brand appears in agent-influenced baskets, how profitability changes when journeys span surfaces, and how value grows as replenishment and cross-sell become automated.

A decade from now, “going to a website to shop” may feel as dated as calling a catalog phone number. Commerce will no longer live in channels. It will be woven into everyday activity and coordinated by AI personal shoppers acting on consumers’ behalf.

In an environment where buying is delegated, the brands that win won’t be the ones with the most channels. They’ll be the ones whose products, data, and economics are easiest for AI agents, and the universal carts they rely on, to understand, assemble, and choose.

*The promo image for this post was generated by AI.