It’s clear that AI agents will fundamentally alter how consumers buy and how companies sell. But no one knows what that transformation will look like—yet.
It could result in a world in which AI agents independently manage your purchases, learning your preferences and completing transactions without you having to view a product page again. Amazon’s Smart Reorders (formerly Dash Replenishment with Alexa) and Instacart’s reordering feature, where resupply happens with minimal human intervention, are precursors of this automation-first path.
Or it could be a market where AI agents serve as intelligent advisors, surfacing all your options and facilitating payment while you retain the final decision rights. Platforms such as ChatGPT and Perplexity have recently launched shopping assistants that guide product discovery and comparison without fully automating the purchasing process.
Or it might be a marketplace where purchases flow through your social networks, with recommendations from friends, influencers, and communities deciding what your AI agents consider. Agents could also amplify the work of creators, professionalizing the business of taste. TikTok Shop and Instagram Shopping’s rapid growth illustrates how discovery, influence, and transactions can merge into a single social stream.
Or it could be a world in which you continue to turn to trusted brands and retailer voices for curation and guidance. L'Oréal Paris’s Beauty Genius, which offers personalized advice 24/7 on its own site, and Amazon’s Rufus, which guides consumers in choosing and purchasing products, show that expert guidance can anchor buying decisions even in an AI age.
All these futures and more are plausible, and some will coexist. Even so, which ones will dominate, and when, remains unclear. What is certain is that the rules governing how consumers discover, evaluate, trust, decide, and buy are being rewritten. This transformation will be existential for marketing. The traditional discipline built around capturing consumer attention, building brand preference, and influencing buying decisions will now have to account for an algorithmic intermediary that may never show a marketing team’s carefully crafted message to a human being.
Moreover, AI adoption is likely to be uneven across product categories, countries, and consumer segments. Consumer behavior will shift fluidly across contexts based on convenience, time, and the stakes involved in each purchase. And agentic technology is evolving faster than planning cycles. Marketing will still be about meeting consumers where they are, but where that will be is difficult to tell.
Success will require marketing strategies that work across agentic scenarios, such as building machine-readable product data and brand signals that algorithms can assess, ensuring accessibility wherever agents look, and understanding how autonomous intermediaries evaluate products. Two imperatives will determine whether brands win or lose: discoverability, the ability to be found by the agents that mediate discovery, and desirability, the power to be wanted by the consumers that those agents serve.
Although most companies are still wrestling with fragmented data, siloed content, and operating models that were built for human-led journeys, CEOs cannot afford to wait any longer for clarity. They must prepare for multiple agentic worlds now, because the price of indecision compounds faster than the cost of making a mistake. Indeed, the marketing challenge isn’t predicting which future will emerge, but building capabilities that are robust enough to succeed regardless of which futures unfold.
Stay ahead with BCG insights on marketing and sales
Why Scenarios Score Over Forecasts
Already, AI search is shifting traffic away from direct e-commerce even as conversational commerce is reinventing the online experience. Even so, most companies are planning for the future based on forecasts that address questions such as: When will agent-driven purchases achieve critical mass? Which product categories will see autonomous purchasing first? When will brand building become less effective than algorithmic optimization?
But those are the wrong questions to ask. The right one is: How do we build capabilities that work whether the market is 10% agent-driven or 90%? Preparedness, not prediction, creates advantage.
When the underlying technology, consumer adoption patterns, and competitive responses are all moving targets, forecasts provide a false sense of precision. Scenario planning, on the other hand, acknowledges what we cannot know today and builds capabilities that will work whether AI agents become dominant intermediaries or remain niche tools. It forces marketers to identify which investments will pay off across possibilities, not just the one that they hope will materialize.
Marketing’s agentic future will be shaped by forces operating at three levels. Each will evolve according to its own timeline and influence the others in tangible ways. And together, they will create a complex, hard-to-predict agentic AI market.
- At the technology layer, the speed at which agentic AI capabilities advance will interact with the dynamics of data flows between brands, retailers, and platforms to determine what agents can do and which players can do it.
- At the human layer, consumer adoption rates, trust in AI, and social influence will determine whether people embrace algorithmic intermediaries or continue to rely on community recommendations and brand relationships.
- At the structural layer, regulatory frameworks will interact with economic conditions and competitive dynamics to determine whether the market consolidates around a few platforms or fragments across many agents, and whose interests AI agents serve.
Because the existence of so many variables complicates decision making, we developed scenarios to identify those that would generate the most distinct futures. (See Exhibit 1.)
We explored which uncertainties would frame a set of mutually exclusive worlds, where optimizing for one would weaken a company’s position in the others. Two variables met this test.
Where will influence reside? Will it lie with social networks and human judgment? Or with algorithmic optimization and data? At the extremes, this will determine what companies must optimize for. If humans drive purchase decisions, success will require investing in brand equity, community relationships, and persuasive storytelling. If algorithms drive them, success will involve focusing on verifiable data, machine-readable attributes, and API accessibility.
How will market power be concentrated? Will it be consolidated among a few dominant platforms? Or will it be distributed across many competing agents? This will decide who controls customer access and where negotiating power lies. In concentrated markets, companies must secure relationships with platform gatekeepers, who control the interface between brands and consumers. But in distributed markets, companies can compete directly for agents’ consideration in more open ecosystems.
Four Agentic AI Futures for Retail
These two dimensions will determine what companies optimize for. The variables are independent of each other: human influence can dominate in either consolidated or distributed markets, as can algorithmic optimization. Different combinations of these variables create four distinct agentic AI landscapes, each with its own rules for winning.(See Exhibit 2.)
The Open Agentic Bazaar
The bazaar is an open, distributed ecosystem where no single platform dominates. Shopping agents browse and transact freely across brands and retailers, moving seamlessly between platforms without friction. However, regional regulations create uneven standards for data sharing, transparency requirements, and agent behavior, forcing companies to navigate a patchwork of compliance regimes.
Brands invest in optimization for each autonomous agent to influence how they evaluate and recommend products. Retailers evolve into network hubs, providing rich product data, real-time inventory, and brand information that agents can query efficiently. Social and creator commerce remain influential, so brands that understand how to feed agent models with community signals and preferences will gain an edge.
Brand Resurgence Through Data Fortresses
In this scenario, a few large brand and retail platforms dominate the market, uniting search, commerce, data, loyalty, and their own agent engines in closed ecosystems that control shopper journeys. Autonomous third-party agents remain marginal; consumers default to brand-led and retailer-driven experiences that offer convenience and familiarity. Social commerce becomes an extension of brand- and retailer-controlled media systems that control recommendations.
Companies compete for visibility within these walled gardens through retail media-paid placements, platform-specific marketing programs, and asymmetric data partnerships, where they share customer insights but receive limited intelligence in return. Success depends on securing favorable terms with the dominant platforms and optimizing for each ecosystem’s proprietary recommendation algorithms.
The Super-App Embrace
The global super-apps launched by a handful of technology giants dominate agentic shopping in this future landscape. They embed AI agents into daily life through voice, augmented reality, and home devices that manage entire shopping journeys. Consumers simply state their needs; the super-app’s agent executes across retailers’ fulfillment systems, optimizing for price, convenience, and loyalty.
Tech platforms own customer relationships, retailers own the warehouse, and brands pay both for access. Brands must navigate technology platforms’ gatekeepers for visibility and access to the consumers inside each super-app’s closed agent layer. Retailers are excluded from customer relationships and become backend logistics providers that compete on fulfillment speed and cost.
A Creator-Led Authenticity Revival
In this scenario, consumers crave connections over algorithmic control and forge closer relationships with creators and communities they trust. Discovery and purchase decisions flow through human networks, which are trusted over AI. No platform dominates; instead, creators maintain independence through multi-platform presences and direct-to-community tools.
Brands rebuild around creator partnerships, transparent storytelling, and decentralized social commerce networks. AI will serve as a tool that helps produce content, manage communities, and personalize recommendations, but humans remain the decision makers and trust anchors. Regulations requiring AI disclosures reinforce consumer preferences for human curation. Data portability rules enable creators to move freely between platforms, preventing any player from monopolizing consumer relationships.
Marketing’s Twin Imperatives for Agentic Worlds
While these scenarios describe dramatically different agentic futures, discoverability and desirability cut across all of them. (See Exhibit 3.)
Discoverability ensures that your brands surface whether agents are browsing open marketplaces, navigating platform algorithms, or filtering through creator recommendations. In a world where consumers may never see a retailer’s website or advertising, being findable by the systems that mediate discovery becomes a prerequisite for relevance. This requires machine-readable product data, participation in the right networks, and optimization of the discovery mechanisms across answer engines, search, and creator ecosystems that dominate each scenario.
Desirability requires building brand strength that commands attention and preference regardless of the interface. Agents don’t shop in a vacuum; they optimize for users’ preferences, values, and behavior. Strong brands with clear differentiation, verified quality, and authentic positioning will enjoy an advantage whether an algorithm is evaluating specifications or a creator is making recommendations. Even in an agentic world, branding won’t matter less—it will matter differently, but even more than it used to.
How discoverability and desirability are used in an agentic AI world will depend on the roles companies play in the marketing value chain, be they brand builders, retailers, or marketing agencies. Moreover, each move must be tailored to the product category, its competitive dynamics, and the company’s scale of operations.
No-Regret Foundational Strategies for Any Agentic Future
Based on our extensive analysis of AI’s role in marketing and retail, we outline several key strategies that will be be key to navigating the world ahead.
Brand equity will become increasingly critical for discoverability and desirability. As a recent BCG study showed, 76% of marketers say that cutting brand spending has a greater adverse impact today than it did five years ago. Their companies focus on building brand equity by using precision tactics such as platform analytics and behavioral insights to capture attention in increasingly noisy markets.
Smart companies anchor trust by aligning campaigns with consumer needs, knowing that trust correlates with an 8 percentage-point boost in First Fast Response, BCG’s predictive metric for future sales. They also measure brand impact with the same financial rigor as conversion, triangulating across methodologies to prove to themselves that brand investments deliver returns.
Answer Engine Optimization (AEO) will be essential for discoverability. Consumers are increasingly turning to generative AI (GenAI) and social media as search tools. While 40% of Gen Z starts its searches on Instagram or Tik Tok, ChatGPT ranked fifth in monthly website visitors in 2025, surpassing Amazon, while Google’s AI Overviews appeared in 21% of all searches.
Because AI answer engines evaluate content differently than humans, marketers must optimize for how they read content, creating clear, context-rich information with proper schema markup. They must maintain a consistent presence across third-party sources such as Reddit and Wikipedia, the most-cited domains in ChatGPT responses, and integrate directly with AI platforms through product feeds and API connections. A recent BCG analysis showed only an 8% to 12% overlap between traditional search results and AI-generated answers, so companies will need both: SEO will capture bottom-funnel intent while AEO can influence top- and middle-funnel agentic AI discovery.
Marketing speed will rise in importance in fast-moving agentic worlds. This makes the shift to an AI-first marketing organization—and the necessary investments in data, technology, and operating models—a key enabler. This change can triple marketing ROI, speed, and volume, translating to 5% to 10% incremental growth and a 15% to 20% increase in efficiency, according to another BCG study.
Marketers can capture real-time signals for microsegments, deploy multi-agent platforms that process data at machine speed, and use synthetic consumers to test messaging across personas and markets. Success will require five capabilities: embedding AI talent in marketing teams, continuous training, clear KPIs, integrated tooling, and governance guardrails. In a world where AI agents mediate discovery and purchase, marketing will have to match consumers by developing its own agentic capabilities.
The Big Bets That Can Set the Stage for Agentic Success
The transition to an agentic marketplace will require making a small number of big bets with large financial commitments. Those decisions will shape how a company competes, how it interfaces with platforms and agents, and where it captures value. Strategies will differ by player, and the returns will depend on the evolution of the agentic market.
Brands
An agentic world offers brands several options. (See Exhibit 4.) One is to build proprietary AI assistants that strengthen consumer relationships on their own platforms. Nike’s NikeAI Beta allows customers to describe what they need in natural language—such as their sport, body type, and goals—and receive a tailored product recommendation through dialogue rather than keyword search. At Starbucks, the Deep Brew app learns each loyalty member’s order history, time-of-day preferences, and even local weather to serve up personalized drink and food recommendations each time they open it. And BMW’s Intelligent Personal Assistant gets to know how its owner drives and lives, adjusting vehicle settings, answering questions, and anticipating needs both through the MyBMW app and the car itself. These represent the growing trend of brands investing in proprietary AI to own the customer relationship on home ground, without ceding the interface to a third party.
Another tactic is to meet consumers inside the AI platforms where they spend time. This approach has been pioneered by aggregators. Expedia and Kayak offer booking through ChatGPT plugins, while OpenTable handles restaurant reservations without the consumer ever leaving the AI platform. That inverts current strategy: instead of pulling consumers to a brand’s channels, the aggregator meets them in their existing workflows. The result is deeper embedding, higher engagement, and potentially lower acquisition costs. What’s counterintuitive is that building inside another company’s ecosystem, without controlling the platform or interface, can deliver more value than maintaining independence.
A third path is to integrate with LLM platforms to reduce reliance on retail intermediaries, enabling discovery and checkout within the former’s conversational interfaces. Fashion and beauty brands such as Glossier, SKIMS, Spanx, and Vuori are beginning to enable direct discovery and transaction flows through ChatGPT without requiring retail intermediaries. Large numbers of Shopify merchants and Etsy sellers are gaining similar capabilities through competing technical standards such as the Agentic Commerce Protocol, co-developed by OpenAI and Stripe, and the Universal Commerce Protocol, backed by Google, Shopify, and over 20 other partners. For brands that have spent decades negotiating with retailers, this represents a structural shift from retail dependence to platform dependence.
Marketing Agencies
Agencies must use this moment to build expertise in agentic systems. The risk is not just lagging in AI adoption, but optimizing for the wrong scenario. Because no one knows whether agentic markets will consolidate around a handful of dominant platforms or fragment across interoperable agents, marketing agencies must invest in capabilities that remain valuable across a range of possible outcomes.
In distributed markets, agencies will need fluency in structured data, API integration, and agent optimization. In consolidated ecosystems, they will need creative excellence combined with the institutional capability to navigate platform gatekeepers. Some clients will prioritize technical integration; others will require community building and trust. The marketing agencies that thrive will not bet on one future, but design operating models that perform across several agentic futures.
For example, WPP has launched WPP Open Pro, giving marketers direct access to its AI tools while partnering with Google on creative automation through a $400-million deal which seeks to embed AI in the operating spine of client work. Omnicom has expanded its Omni operating system, deploying GenAI assistants to accelerate planning and optimization as well as agents to simulate consumer response. These moves reflect a shift from campaign execution to infrastructure, interoperability, and real-time decision systems.
Retailers
In an agentic world, retailers must decide whether they want to play the destination game or the evaluation game. Destination retailers will aim to remain the consumer’s first port of call, where intent is formed and loyalty reinforced. Evaluation retailers will accept that discovery may happen elsewhere and compete to win when agents compare alternatives.
If a retailer chooses the destination strategy, it must invest in owning ecosystems with proprietary apps, seamless in-store integration, memberships, private labels, and loyalty architectures that embed the retailer into daily routines. Target has pursued this path, building its Circle loyalty program, tiered membership, private labels, and same-day delivery through Shipt into a single integrated system. It is designed to make Target the default choice before a purchase decision is made. This approach works best where purchase frequency is high and differentiation is meaningful, conditions that allow habit and data to compound over time.
By comparison, a retailer that chooses the evaluation strategy must optimize for machine-mediated comparison: structured product data, transparent pricing, fulfillment reliability, verified reviews, and API interoperability across platforms and agentic interfaces. Wayfair has built on these foundations, competing not through ecosystem lock-in but through the depth and quality of its product catalog, pricing transparency, and fulfillment reliability. According to a BCG study, this model is strongest where retailers can win on utility by offering the lowest price, the fastest delivery, or the best product fit for a specific consumer need.
Retailers don’t have to abandon one model for another, but they may not be able to optimize equally for both. Destination dominance requires long-horizon ecosystem investment; evaluation dominance demands operational precision and machine visibility. The danger is taking half-measures by investing in both without committing to either.
As the agentic future dawns, the smart move is not to pick a single world. It is to prepare for all four scenarios: the open bazaar where agents roam freely, the walled gardens where retailers reign, the super-app world where tech giants own the customer, and the creator revival where human trust takes precedence over algorithmic precision. For each scenario, CEOs should ask what the threats and opportunities are for their business, which capabilities would make them competitive, and what they would do differently if they knew that world was coming.
The signals that reveal which future is arriving are already visible in the growing share of purchase journeys that begin inside AI agents rather than search engines, in whether AI assistants recommend brands by name or by specification, and in how consumers feel about content they know was generated by machines. CEOs who read the signals early, build the capabilities that travel across futures, and place their big bets with conviction will not just adapt their organizations to the agentic future. They will shape it—and in doing so, write the rules of marketing for an AI world that rivals will be forced to follow.
