AI Will Buy Your Groceries: But Someone Else Profits
The future of shopping isn't happening on a screen. It's happening in the spaces between your words and a machine's understanding of what you want. When you ask an AI agent to find you the best noise-cancelling headphones under £300, you're not just outsourcing a Google search. You're delegating an entire decision-making process to an algorithmic intermediary that will reshape how billions of pounds flow through the digital economy.
This is agentic commerce: AI systems that browse, compare, negotiate, and purchase on behalf of humans. And it's already here. OpenAI's ChatGPT now offers instant checkout for purchases from over one million Shopify merchants. Perplexity launched its Comet browser with AI agents that can autonomously complete purchases from any retailer. Opera introduced Browser Operator, the first major browser with AI-based agentic capabilities built directly into its architecture. Google is expanding its AI Mode shopping interface across the United States, adding capabilities that let customers track prices and confirm purchases without ever visiting a retailer's website.
The numbers tell a story of exponential transformation. Traffic to US retail sites from generative AI browsers and chat services increased 4,700 per cent year-over-year in July 2025, according to industry tracking data. McKinsey projects that by 2030, the US business-to-consumer retail market alone could see up to one trillion dollars in orchestrated revenue from agentic commerce, with global projections reaching three trillion to five trillion dollars.
But these astronomical figures obscure a more fundamental question: When AI agents become the primary interface between consumers and commerce, who actually benefits? The answer is forcing a reckoning across the entire e-commerce ecosystem, from multinational retailers to affiliate marketers, from advertising platforms to regulatory bodies. Because agentic commerce doesn't just change how people shop. It fundamentally rewrites the rules about who gets paid, who gets seen, and who gets trusted in the digital marketplace.
The Funnel Collapses
The traditional e-commerce funnel has been the foundational model of online retail for two decades. Awareness leads to interest, interest leads to consideration, consideration leads to conversion. Each stage represented an opportunity for merchants to influence behaviour through advertising, product placement, personalised recommendations, and carefully optimised user experience. The funnel existed because friction existed: the cognitive load of comparing options, the time cost of browsing multiple sites, the effort required to complete a transaction.
AI agents eliminate that friction by compressing the entire funnel into a single conversational exchange. When a customer arriving via an AI agent reaches a retailer's site, they're already further down the sales funnel with stronger intent to purchase. Research shows these customers are ten per cent more engaged than traditional visitors. The agent has already filtered options, evaluated trade-offs, and narrowed the field. The customer isn't browsing. They're buying.
This compression creates a paradox for retailers. Higher conversion rates and more qualified traffic represent the holy grail of e-commerce optimisation. Yet if the AI agent can compress browsing, selection, and checkout into the same dialogue, retailers that sit outside the conversation risk ceding both visibility and sales entirely.
Boston Consulting Group's modelling suggests potential earnings before interest and taxes erosion of up to 500 basis points for retailers, stemming from price transparency, smaller order sizes, and agent platform fees. That five per cent margin compression might not sound catastrophic until you consider that many retailers operate on margins of ten to fifteen per cent. Agentic commerce could eliminate a third of their profitability.
The risks extend beyond margins. Retailers face diminished direct access to customers, weaker brand loyalty, and growing dependence on intermediary platforms. When customers interact primarily with an AI agent rather than a retailer's website or app, the retailer loses the ability to shape the shopping experience, collect first-party data, or build lasting relationships. The brand becomes commoditised: a product specification in an agent's database rather than a destination in its own right.
This isn't speculation. Walmart announced a partnership with OpenAI enabling seamless “chat to checkout” experiences. Shopify integrated with ChatGPT to allow instant purchases from its merchant base. Etsy followed suit. These aren't defensive moves. They're admissions that the platform layer is shifting, and retailers must establish presence where the conversations are happening, even if it means surrendering control over the customer relationship.
The Revenue Model Revolution
If agentic commerce destroys the traditional funnel, it also demolishes the advertising models built upon that funnel. Consider Google Shopping, which has operated for years on a cost-per-click model with effective commission rates around twelve per cent. Or Amazon, whose marketplace charges sellers approximately fifteen per cent in fees and generates billions more through advertising within search results and product pages. These models depend on human eyeballs viewing sponsored listings, clicking through to product pages, and making purchase decisions influenced by paid placement.
AI agents have no eyeballs. They don't see banner ads or sponsored listings. They process structured data, evaluate parameters, and optimise for the objectives their users specify. The entire edifice of digital retail advertising, which represents a 136 billion dollar industry in 2025, suddenly faces an existential question: How do you advertise to an algorithm?
The early answer appears to be: You don't advertise. You pay for performance. OpenAI has reportedly discussed a two per cent affiliate commission model for purchases made through its shopping features. That's six times lower than Google Shopping's traditional rates and seven times lower than Amazon's marketplace fees. The economics are straightforward. In a world where AI agents handle product discovery and comparison, platforms can charge lower fees because they're not operating expensive advertising infrastructure or maintaining complex seller marketplaces. They're simply connecting buyers and sellers, then taking a cut of completed transactions.
This shift from advertising-based revenue to performance-based commissions has profound implications. Advertisers will spend 12.42 billion dollars on affiliate programmes in 2025, up 10.2 per cent year-over-year, driving thirteen per cent of US e-commerce sales. The affiliate marketing ecosystem has adapted quickly to the rise of AI shopping agents, with seventy per cent of citations for some retailers in large language models stemming from affiliate content.
But the transition hasn't been smooth. Retail affiliate marketing revenues took a hit of over fifteen per cent year-over-year in the second quarter of 2024, when Google's search algorithm updates deprioritised many affiliate sites. If ChatGPT or Perplexity become the primary shopping interfaces, and those platforms negotiate direct relationships with merchants rather than relying on affiliate intermediaries, the affiliate model could face an existential threat.
Yet the performance-based structure of affiliate marketing may also be its salvation. Cost-per-acquisition and revenue-share pricing align perfectly with agentic commerce, where marketing dollars are spent only when a purchase is made. Industry analysts predict retail media networks will reshape into affiliate-like ecosystems, complete with new metrics such as “cost per agent conversion.”
The retail media network model faces even more severe disruption. Retail media networks, which allow brands to advertise within retailer websites and apps, are projected to reach 136 billion dollars in value during 2025. But these networks depend on high human traffic volumes consuming brand messages, sponsored product listings, and targeted advertisements. When agentic AI threatens those traffic volumes by handling shopping outside retailer environments, the entire business model begins to crumble.
The industry response has been to pivot from business-to-consumer advertising to what executives are calling business-to-AI: competing for algorithmic attention rather than human attention. Traditional brand building, with its emphasis on emotional connections, lifestyle aspirations, and community, suddenly becomes the most valuable marketing strategy. Because whilst AI agents can evaluate specifications and compare prices, they still rely on the corpus of available information to make recommendations. A brand that has invested in thought leadership, earned media coverage, and authentic community engagement will appear more frequently in that corpus than a brand that exists only as a product listing in a database.
The new battleground isn't the moment of purchase. It's the moment of instruction, when a human tells an AI agent what they're looking for. Influence that initial framing and you influence the entire transaction.
The Merchant's Dilemma
For retailers, agentic commerce presents an agonising choice. Participate in these new platforms and surrender margin, control, and customer data. Refuse to participate and risk becoming invisible to a growing segment of high-intent shoppers.
The mathematics of merchant incentives in this environment grow complex quickly. If Target and Walmart stock the same product at the same price, how does an AI agent decide which retailer to recommend? In traditional e-commerce, the answer involves search engine optimisation, paid advertising, customer reviews, shipping speed, and loyalty programme benefits. In agentic commerce, the answer increasingly depends on which merchant is willing to pay the AI platform a performance incentive.
Industry analysts worry this creates a “pay to play” dynamic reminiscent of Google's shift from organic search results to advertising-dominated listings. Anyone who has used Google knows how much the first page of search results is stuffed with sponsored listings. Could agentic commerce go the same way? Currently, executives at AI companies insist their algorithms pick the best possible choices without pay-to-play arrangements. But when significant money is involved, the concern is whether those principles can hold.
Perplexity has directly criticised Amazon for being “more interested in serving you ads, sponsored results, and influencing your purchasing decisions with upsells and confusing offers.” The criticism isn't just rhetorical posturing. It's a competitive claim: that AI agents provide a cleaner, more consumer-focused shopping experience precisely because they're not corrupted by advertising revenue. Whether that purity can survive as agentic commerce scales to trillions of pounds in transaction volume remains an open question.
Some merchants are exploring alternative incentive structures. Sales performance incentive funds, where retailers pay commissions to AI platforms only when purchases are completed, align merchant interests with platform performance. Dynamic pricing strategies, where retailers offer AI platforms exclusive pricing in exchange for preferential recommendations, create a more transparent marketplace for algorithmic attention. Subscription models, where merchants pay fixed fees for inclusion in AI agent recommendation databases, avoid the pay-per-click inflation that has plagued search advertising.
But each of these approaches raises questions about transparency, fairness, and consumer welfare. If an AI agent recommends Target over Walmart because Target pays a higher commission, is that a betrayal of the user's trust? Or is it simply the same economic reality that has always governed retail, now made more efficient through automation? The answer depends largely on disclosure: whether users understand the incentives shaping the recommendations they receive.
The Transparency Crisis
Trust is the currency of AI shopping agents. If users don't trust that an agent is acting in their best interests, they won't delegate purchasing decisions. And trust requires transparency: understanding how recommendations are generated, what incentives influence those recommendations, and whether the agent is optimising for the user's preferences or the platform's profit.
The current state of transparency in AI shopping is, charitably, opaque. Most AI platforms provide little visibility into their recommendation algorithms. Users don't know which merchants have paid for preferential placement, how commissions affect product rankings, or what data is being used to personalise suggestions. The Federal Trade Commission has made clear there is no AI exemption from existing consumer protection laws, and firms deploying AI systems have an obligation to ensure those systems are transparent, explainable, fair, and empirically sound.
But transparency in AI systems is technically challenging. The models underlying ChatGPT, Claude, or Perplexity are “black boxes” even to their creators: neural networks with billions of parameters that produce outputs through processes that defy simple explanation. Algorithmic accountability requires examination of how results are reached, including transparency and justification of the AI model design, setup, and operation. That level of scrutiny is difficult when the systems themselves are proprietary and commercially sensitive.
The FTC has responded by launching Operation AI Comply, taking action against companies that rely on artificial intelligence to supercharge deceptive or unfair conduct. Actions have targeted companies promoting AI tools that enable fake reviews, businesses making unsupported claims about AI capabilities, and platforms that mislead consumers about how AI systems operate. The message is clear: automation doesn't absolve responsibility. If an AI agent makes false claims, deceptive recommendations, or unfair comparisons, the platform operating that agent is liable.
Bias represents another dimension of the transparency challenge. Research on early AI shopping agents revealed troubling patterns. Agents failed to conduct exhaustive comparisons, instead settling for the first “good enough” option they encountered. This creates what researchers call a “first-proposal bias” that gives response speed a ten to thirty times advantage over actual quality. If an agent evaluates the first few results more thoroughly than later results, merchants have an incentive to ensure their products appear early in whatever databases the agent queries.
Data bias, algorithmic bias, and user bias are the main types of bias in AI e-commerce systems. Data bias occurs when training data isn't representative of actual shopping patterns, leading to recommendations that favour certain demographics, price points, or product categories. Algorithmic bias emerges from how models weigh different factors, potentially overvaluing characteristics that correlate with protected categories. User bias happens when AI agents learn from and amplify existing consumer prejudices rather than challenging them.
The automation bias problem compounds these challenges. People may be unduly trusting of answers from machines which seem neutral or impartial. Many chatbots are effectively built to persuade, designed to answer queries in confident language even when those answers are fictional. The tendency to trust AI output creates vulnerability when that output is shaped by undisclosed commercial incentives or reflects biased training data.
Microsoft recently conducted an experiment where they gave AI agents virtual currency and instructed them to make online purchases. The agents spent all the money on scams. This wasn't a failure of the AI's reasoning capability. It was a failure of the AI's ability to assess trust and legitimacy in an environment designed to deceive. If sophisticated AI systems from a leading technology company can be systematically fooled by online fraud, what does that mean for consumer protection when millions of people delegate purchasing decisions to similar agents?
The Regulatory Response
Regulators worldwide are scrambling to develop frameworks for agentic commerce before it becomes too embedded to govern effectively. New AI-specific laws have emerged to mandate proactive transparency, bias prevention, and consumer disclosures not otherwise required under baseline consumer protection statutes.
The FTC's position emphasises that existing consumer protection laws apply to AI systems. Using artificial intelligence and algorithms doesn't provide exemption from legal obligations around truthfulness, fairness, and non-discrimination. The agency has published guidance stating that AI tools should be transparent, explainable, fair, and empirically sound, whilst fostering accountability.
European regulators are taking a more prescriptive approach through the AI Act, which classifies AI systems by risk level and imposes requirements accordingly. Shopping agents that significantly influence purchasing decisions would likely qualify as high-risk systems, triggering obligations around transparency, human oversight, and impact assessment. The regulation mandates clear disclosure of whether an entity is human or artificial, responding to the increasing sophistication of AI interactions. Under the AI Act's framework, providers of high-risk AI systems must maintain detailed documentation of their training data, conduct conformity assessments before deployment, and implement post-market monitoring to detect emerging risks. Violations can result in fines up to seven per cent of global annual turnover.
But enforcement remains challenging. The opacity of black box models means consumers have no transparency into how exactly decisions are being made. Regulators often lack the technical expertise to evaluate these systems, and by the time they develop that expertise, the technology has evolved. The European Union is establishing an AI Office with dedicated technical staff and budget to build regulatory capacity, whilst the UK is pursuing a sector-specific approach that empowers existing regulators like the Competition and Markets Authority to address AI-related harms in their domains.
The cross-border nature of AI platforms creates additional complications. An AI agent operated by a US company, trained on data from multiple countries, making purchases from international merchants, creates a jurisdictional nightmare. Which country's consumer protection laws apply? Whose privacy regulations govern the data collection? Who has enforcement authority when harm occurs? The fragmentation extends beyond Western democracies. China's Personal Information Protection Law and algorithmic recommendation regulations impose requirements on AI systems operating within its borders, creating a third major regulatory regime that global platforms must navigate.
Industry self-regulation has emerged to fill some gaps. OpenAI and Anthropic developed the Agentic Commerce Protocol, a technical standard for how AI agents should interact with merchant systems. The protocol includes provisions for identifying agent traffic, disclosing commercial relationships, and maintaining transaction records. Google and Amazon rely on separate, incompatible systems, making it difficult for merchants to translate product catalogues into multiple formats.
The question of liability looms large. When an AI agent makes a purchase that the user later regrets, who is responsible? The user who gave the instruction? The platform that operated the agent? The merchant that fulfilled the order? Traditional consumer protection frameworks assume human decision-makers at each step. Agentic commerce distributes decision-making across human-AI interactions in ways that blur lines of responsibility.
The intellectual property dimensions add further complexity. Amazon has sued Perplexity, accusing the startup of violating its terms of service by using AI agents to access the platform without disclosing their automated nature. Amazon argues that Perplexity's agents degrade the Amazon shopping experience by showing products that don't incorporate personalised recommendations and may not reflect the fastest delivery options available. Perplexity counters that since its agent acts on behalf of a human user's direction, the agent automatically has the same permissions as the human user.
This dispute encapsulates the broader regulatory challenge: existing legal frameworks weren't designed for a world where software agents act autonomously on behalf of humans, making decisions, negotiating terms, and executing transactions.
The Power Redistribution
Step back from the technical and regulatory complexities, and agentic commerce reveals itself as fundamentally about power. Power to control the shopping interface. Power to influence purchasing decisions. Power to capture transaction fees. Power to shape which businesses thrive and which wither.
For decades, that power has been distributed across an ecosystem of search engines, social media platforms, e-commerce marketplaces, payment processors, and retailers themselves. Google controlled discovery through search. Facebook controlled attention through social feeds. Amazon controlled transactions through its marketplace. Each entity extracted value from its position in the funnel, and merchants paid tribute at multiple stages to reach customers.
Agentic commerce threatens to consolidate that distributed power into whoever operates the AI agents that consumers trust. If ChatGPT becomes the primary shopping interface for hundreds of millions of users, OpenAI captures influence that currently belongs to Google, Amazon, and every retailer's individual website. The company that mediates between consumer intent and commercial transaction holds extraordinary leverage over the entire economy.
This consolidation is already visible in partnership announcements. When Walmart, Shopify, and Etsy all integrate with ChatGPT within weeks of each other, they're acknowledging that OpenAI has become a gatekeeper they cannot afford to ignore. The partnerships are defensive, ensuring presence on a platform that could otherwise bypass them entirely.
But consolidation isn't inevitable. The market could fragment across multiple AI platforms, each with different strengths, biases, and commercial relationships. Google's AI Mode might excel at product discovery for certain categories. Perplexity's approach might appeal to users who value transparency over convenience. Smaller, specialised agents could emerge for specific verticals like fashion, electronics, or groceries.
The trajectory will depend partly on technical factors: which platforms build the most capable agents, integrate with the most merchants, and create the smoothest user experiences. But it will also depend on trust and regulation. If early AI shopping agents generate high-profile failures, consumer confidence could stall adoption. If regulators impose strict requirements that only the largest platforms can meet, consolidation accelerates.
For consumers, the implications are ambiguous. Agentic commerce promises convenience, efficiency, and potentially better deals through automated comparison and negotiation. Customers arriving via AI agents already demonstrate higher engagement and purchase intent. More than half of consumers anticipate using AI assistants for shopping by the end of 2025. Companies deploying AI shopping agents are delivering thirty per cent more conversions and forty per cent faster order fulfilment.
But those benefits come with risks. Loss of serendipity and discovery as agents optimise narrowly for stated preferences rather than exposing users to unexpected products. Erosion of privacy as more shopping behaviour flows through platforms that profile and monetise user data. Concentration of market power in the hands of a few AI companies that control access to billions of customers. Vulnerability to manipulation if agents' recommendations are influenced by undisclosed commercial arrangements.
Consider a concrete scenario. A parent asks an AI agent to find educational toys for a six-year-old who loves science. The agent might efficiently identify age-appropriate chemistry sets and astronomy kits based on thousands of product reviews and educational research. But if the agent prioritises products from merchants paying higher commissions over genuinely superior options, or if it lacks awareness of recent safety recalls, convenience becomes a liability. The parent saves time but potentially compromises on quality or safety in ways they would have caught through traditional research.
Marketplace or Manipulation
Agentic commerce is not a future possibility. It is a present reality growing at exponential rates. The question is not whether AI shopping agents will reshape retail, but how that reshaping will unfold and who will benefit from the transformation.
The optimistic scenario involves healthy competition between multiple AI platforms, strong transparency requirements that help users understand recommendation incentives, effective regulation that prevents the worst abuses whilst allowing innovation, and merchants who adapt by focusing on brand building, product quality, and authentic relationships.
In this scenario, consumers enjoy unprecedented convenience and potentially lower prices through automated comparison shopping. Merchants reach highly qualified customers with strong purchase intent. AI platforms create genuine value by reducing friction and improving matching between needs and products. Regulators establish guardrails that prevent manipulation whilst allowing experimentation. Picture a marketplace where an AI agent negotiates bulk discounts on behalf of a neighbourhood buying group, secures better warranty terms through automated comparison of coverage options, and flags counterfeit products by cross-referencing manufacturer databases, all whilst maintaining transparent logs of its decision-making process that users can audit.
The pessimistic scenario involves consolidation around one or two dominant AI platforms that extract monopoly rents, opaque algorithms shaped by undisclosed commercial relationships that systematically favour paying merchants over best products, regulatory capture or inadequacy that allows harmful practices to persist, and a race to the bottom on merchant margins that destroys business viability for all but the largest players.
In this scenario, consumers face an illusion of choice backed by recommendations shaped more by who pays the AI platform than by genuine product quality. Merchants become commodity suppliers in a system they can't influence without paying increasing fees. AI platforms accumulate extraordinary power and profit through their gatekeeper position. Imagine a future where small businesses cannot afford the fees to appear in AI agent recommendations, where platforms subtly steer purchases toward their own private-label products, and where consumers have no practical way to verify whether they're receiving genuinely optimal recommendations or algorithmically optimised profit extraction.
Reality will likely fall somewhere between these extremes. Some markets will consolidate whilst others fragment. Some AI platforms will maintain rigorous standards whilst others cut corners. Some regulators will successfully enforce transparency whilst others lack resources or authority. The outcome will be determined by choices made over the next few years by technology companies, policymakers, merchants, and consumers themselves.
The Stakeholder Reckoning
For technology companies building AI shopping agents, the critical choice is whether to prioritise short-term revenue maximisation through opaque commercial relationships or long-term trust building through transparency and user alignment. The companies that choose trust will likely capture sustainable competitive advantage as consumers grow more sophisticated about evaluating AI recommendations.
For policymakers, the challenge is crafting regulation that protects consumers without stifling the genuine benefits that agentic commerce can provide. Disclosure requirements, bias auditing, interoperability standards, and clear liability frameworks can establish baseline guardrails without prescribing specific technological approaches. The most effective regulatory strategies will focus on outcomes rather than methods: requiring transparency in how recommendations are generated, mandating disclosure of commercial relationships that influence agent behaviour, establishing accountability when AI systems cause consumer harm, and creating mechanisms for independent auditing of algorithmic decision-making. Policymakers must act quickly enough to prevent entrenchment of harmful practices but thoughtfully enough to avoid crushing innovation that could genuinely benefit consumers.
For merchants, adaptation means shifting from optimising for human eyeballs to optimising for algorithmic evaluation and human trust simultaneously. The retailers that will thrive are those that maintain compelling brands, deliver genuine value, and build direct relationships with customers that no AI intermediary can fully replace. This requires investment in product quality, authentic customer service, and brand building that goes beyond algorithmic gaming. Merchants who compete solely on price or visibility in AI agent recommendations will find themselves in a race to the bottom. Those who create products worth recommending and brands worth trusting will discover that AI agents amplify quality rather than obscuring it.
For consumers, the imperative is developing critical literacy about how AI shopping agents work, what incentives shape their recommendations, and when to trust algorithmic suggestions versus conducting independent research. Blind delegation is dangerous. Thoughtful use of AI as a tool for information gathering and comparison, combined with final human judgment, represents the responsible approach. This means asking questions about how agents generate recommendations, understanding what commercial relationships might influence those recommendations, and maintaining the habit of spot-checking AI suggestions against independent sources. Consumer demand for transparency can shape how these systems develop, but only if consumers actively seek that transparency rather than passively accepting algorithmic guidance.
Who Controls the Algorithm Controls Commerce
The fundamental question agentic commerce poses is who gets to shape the marketplace of the future. Will it be the AI platforms that control the interface? The merchants with the deepest pockets to pay for visibility? The regulators writing the rules? Or the consumers whose aggregate choices ultimately determine what succeeds?
The answer is all of the above, in complex interaction. But that interaction will produce very different outcomes depending on the balance of power. If consumers remain informed and engaged, if regulators act decisively to require transparency, if merchants compete on quality rather than just algorithmic gaming, and if AI platforms choose sustainable trust over exploitative extraction, then agentic commerce could genuinely improve how billions of people meet their needs.
If those conditions don't hold, we're building a shopping future where the invisible hand of the market gets replaced by the invisible hand of the algorithm, and where that algorithm serves the highest bidder rather than the human asking for help. The stakes are not just commercial. They're about what kind of economy we want to inhabit: one where technology amplifies human agency or one where it substitutes algorithmic optimisation for human choice.
The reshape is already underway. The revenue is already flowing through new channels. The questions about trust and transparency are already urgent. What happens next depends on decisions being made right now, in boardrooms and regulatory offices and user interfaces, about how to build the infrastructure of algorithmic commerce. Get those decisions right and we might create something genuinely better than what came before. Get them wrong and we'll spend decades untangling the consequences.
The invisible hand of AI is reaching for your wallet. The question is whether you'll notice before it's already spent your money.
Sources and References
McKinsey & Company (2025). “The agentic commerce opportunity: How AI agents are ushering in a new era for consumers and merchants.” McKinsey QuantumBlack Insights.
Boston Consulting Group (2025). “Agentic Commerce is Redefining Retail: How to Respond.” BCG Publications.
Opera Software (March 2025). “Opera becomes the first major browser with AI-based agentic browsing.” Opera Newsroom Press Release.
Opera Software (May 2025). “Meet Opera Neon, the new AI agentic browser.” Opera News Blog.
Digital Commerce 360 (October 2025). “McKinsey forecasts up to $5 trillion in agentic commerce sales by 2030.”
TechCrunch (September 2025). “OpenAI takes on Google, Amazon with new agentic shopping system.”
TechCrunch (March 2025). “Opera announces a new agentic feature for its browser.”
PYMNTS.com (2025). “Agentic AI Is Quietly Reshaping the eCommerce Funnel.”
Retail Brew (October 2025). “AI agents are becoming a major e-commerce channel. Will retailers beat them or join them?”
eMarketer (2025). “As consumers turn to AI for shopping, affiliate marketing is forging its own path.”
Retail TouchPoints (2025). “Agentic Commerce Meets Retail ROI: How the Affiliate Model Powers the Future of AI-Led Shopping.”
Federal Trade Commission (2023). “The Luring Test: AI and the engineering of consumer trust.”
Federal Trade Commission (2025). “AI and the Risk of Consumer Harm.”
Federal Trade Commission (2024). “FTC Announces Crackdown on Deceptive AI Claims and Schemes.”
Bloomberg (November 2025). “Amazon Demands Perplexity Stop AI Tool's Purchasing Ability.”
CNBC (November 2025). “Perplexity AI accuses Amazon of bullying with legal threat over Comet browser.”
Retail Dive (November 2025). “Amazon sues Perplexity over AI shopping agents.”
Criteo (2025). “Retail media in the agentic era.”
Bizcommunity (2025). “Retail media: Agentic AI commerce arrives, estimated value of $136bn in 2025.”
The Drum (June 2025). “How AI is already innovating retail media's next phase.”
Brookings Institution (2024). “Algorithmic bias detection and mitigation: Best practices and policies to reduce consumer harms.”
Lawfare Media (2024). “Are Existing Consumer Protections Enough for AI?”
The Regulatory Review (2025). “A Modern Consumer Bill of Rights in the Age of AI.”
Decrypt (November 2025). “Microsoft Gave AI Agents Fake Money to Buy Things Online. They Spent It All on Scams.”
Mastercard (April 2025). “Mastercard unveils Agent Pay, pioneering agentic payments technology to power commerce in the age of AI.”
Payments Dive (2025). “Visa, Mastercard race to agentic AI commerce.”
Fortune (October 2025). “Walmart's deal with ChatGPT should worry every ecommerce small business.”
Harvard Business Review (February 2025). “AI Agents Are Changing How People Shop. Here's What That Means for Brands.”
Adweek (2025). “AI Shopping Is Here but Brands and Retailers Are Still on the Sidelines.”
Klaviyo Blog (2025). “AI Shopping: 6 Ways Brands Can Adapt Their Online Presence.”
Tim Green UK-based Systems Theorist & Independent Technology Writer
Tim explores the intersections of artificial intelligence, decentralised cognition, and posthuman ethics. His work, published at smarterarticles.co.uk, challenges dominant narratives of technological progress while proposing interdisciplinary frameworks for collective intelligence and digital stewardship.
His writing has been featured on Ground News and shared by independent researchers across both academic and technological communities.
ORCID: 0009-0002-0156-9795 Email: tim@smarterarticles.co.uk
#HumanInTheLoop #AgenticCommerce #DigitalEconomy #AlgorithmicTrust
