The proxy procurement revolution: How generative artificial intelligence redefines e-commerce

Generative AI and Agentic Commerce: Restructuring E-Commerce from Automation to Value Redistribution

Generative AI represents a qualitative leap in e-commerce, going beyond merely improving user experience to redefining the structure of the market itself. Field research indicates that generative AI applications can increase sales by up to 16.3%, with a 21.7% lift in purchase conversion rates—without any increase in operational costs or labor. At the same time, the Agentic Commerce model is emerging as a restructuring force, with AI agents taking on search, comparison, and purchasing tasks on behalf of users, threatening to redistribute value across the entire e-commerce chain. This article examines the quantitative and qualitative dimensions of this transformation, analyzing the enormous opportunities it creates and the structural risks it imposes on brands and retailers.

First: Generative AI as a Productivity Tool — Measurable Field Results

1. Real Figures from the Largest E-Commerce Market

In a field study covering millions of users on a global e-commerce platform, researchers from Columbia Business School tested seven generative AI applications between September 2023 and June 2024. The results were as follows:

This tangible impact reflects gains in Total Factor Productivity, where output rose without an increase in labor or capital inputs. AI not only accelerated work but made the market smoother, raising conversion rates without affecting average basket value. Collectively, this translates into an additional annual value of $4.6 to $5 per consumer—a figure of economic significance given the platform’s global scale and the early adoption stage of the technologies.

2. Fair Distribution of Benefits

Unlike previous technological shifts that widened gaps, the study showed that generative AI acts as a “market equalizer.” Smaller and less experienced sellers benefited more than established large merchants, and less active consumers (such as new or less experienced platform users) benefited more than their seasoned counterparts. For example, in the chatbot experiment, conversion rates among inexperienced users jumped by 26% versus 20% for veteran shoppers. The reason is that AI bridged gaps in information, experience, and access, helping both sides of the market catch up. Similarly, an IEEE academic study demonstrated the potential of using a generative AI framework to create personalized advertising content, resulting in a notable rise in click-through rates and user loyalty.

Second: Agentic Commerce — Redefining the Entire Purchase Journey

1. From Search to Answer — A Radical Shift in Entry Points

Industry experts expect 2026 to bring a “structural reckoning” for e-commerce, as product discovery shifts from traditional search engines and store listings to conversational tools and smart marketplaces. The purchase journey now begins with a “question,” not a “search.” Users ask their agents things like: “Find me a birthday gift for a hiking enthusiast under $100” or “Reorder my weekly groceries with loyalty discounts.” The automated agent reviews past orders, checks current inventory, applies discounts, and completes the order while keeping the user informed. This model, supported by protocols such as the Agentic Commerce Protocol (ACP), is seen as a qualitative leap comparable to the emergence of e-commerce marketplaces themselves.

2. The “Supermarket” of Product Discovery

AI platforms have emerged as new gateways to purchase, each playing a different role in the consumer journey:

• OpenAI’s ChatGPT: Has the broadest user base and, through partnerships with Etsy, Shopify, and Walmart, is moving from offering advice to completing transactions.

• Google Gemini: Relies on Google’s search and shopping data and integrates automated payment into the surfaces users are accustomed to using for product discovery.

• Perplexity: Stands out as a research tool for purchases requiring in-depth comparisons, with integration with PayPal to facilitate payment within the conversation.

• Amazon Rufus: Operates within the Amazon app, with first-class access to the catalog, reviews, and the Prime logistics network, enabling features such as “Buy for Me” and “automatic purchasing.”

3. Amazon — The Most Radical Restructuring Model

Through its “Buy for Me” feature, powered by Amazon Nova and Anthropic’s Claude, Amazon has expanded its interface to include other retailers. A user can find a product on a third-party site, and Amazon’s agent navigates to that site on their behalf, adds the product to the cart, enters the shipping and payment details stored with Amazon, and completes the purchase—all without leaving the Amazon app. This model turns Amazon into a unified interface for the entire purchase journey, while retailers risk becoming just a link in the supply chain, losing their direct relationship with customers and their valuable data. Estimates suggest the agentic commerce market could reach $236 billion by 2034.

Third: Consumer Gains and Retailer Risks

1. Clear Consumer Gains

• Time and effort savings: Experiments such as the Albertsons AI grocery app showed it could cut shopping time from 46 minutes to just 4 minutes.

• Money savings: Early evidence indicates that Prime users of the “automatic purchasing” feature (which monitors prices and executes deals when the target price is reached) saved an average of 20% per purchase.

• Hyper-personalization: Agents can learn from purchase history and preferences to deliver highly accurate, personalized recommendations.

2. Existential Risks for Retailers and Brands

• Loss of control over the interface: When a product is displayed within an AI conversation, the brand loses control over how it is presented, its context, and its ability to persuade the customer or upsell additional products.

• Loss of data: The flow of customer and purchasing behavior data is the price merchants pay for access to these new channels. This threatens their retail, targeting, and customer retention capabilities.

• Regulatory and technical complexity: Brands face the challenge of adapting to multiple, incompatible protocols (such as OpenAI and Anthropic’s ACP versus Google’s and Amazon’s proprietary systems), creating fragmentation. PayPal noted that building direct integrations with every platform is impractical, highlighting the role of intermediary infrastructure.

Fourth: The Adaptation Path — From Content Optimization to Agent Governance

1. Generative Experience Optimization (GEO)

The shift from Search Engine Optimization (SEO) to Generative Experience Optimization (GEO) is a fundamental transformation. Instead of optimizing keywords and links, brands must now optimize their data to be “machine-readable” by AI agents. This requires:

• Structured, accurate product data: Providing precise details on specifications, compatibility, certifications, and dimensions—the backbone agents rely on for filtering and recommendation.

• Context-rich content: Detailed descriptions, reviews, images, and videos so the agent understands the product’s potential uses.

2. Strategies for Dealing with Agents

Six practical guidelines have been proposed for retailers to navigate this new era:

1. Invest in GEO: Make the catalog “recommendable” by agents through structured data, rich reviews, and real-time price and inventory availability.

2. Build an in-store chatbot assistant: To offer a similar experience and maintain a direct relationship with the customer.

3. Open a commerce API: To allow platforms like ChatGPT and Gemini to handle transactions directly.

4. Add “agent-friendly” retention tools: Such as one-click reordering, automatic replenishment, and price-drop alerts.

5. Track agent traffic: As an independent channel with its own economics.

6. Govern brand representation: Monitor how external agents present the brand, its products, and prices.

Fifth: Conclusion and Future Outlook

Generative AI is redefining e-commerce on three levels:

1. Automation level: Automating routine tasks such as content creation and search optimization, yielding measurable productivity gains without increased costs.

2. Experience level: Transforming the purchase journey from browsing listings to a personalized conversation, leveraging deep understanding of context and intent.

3. Market structure level: Redistributing value by disintermediating traditional middlemen, as platforms (Amazon, Google, OpenAI) compete to become the single purchase gateway, threatening the role of retailers and brands.

The message is clear: the future belongs not only to those who use AI, but to those who recognize that the battle has shifted from optimizing web pages to presence and trust in automated conversations. Success requires a radical transformation in data, product, and governance strategies—where content becomes “machine-readable” and the brand becomes “trustworthy” in the eyes of digital agents.

Sources and References

1. Columbia Business School. (2025). Generative AI Is Quietly Supercharging Online Retail.

2. Yahoo Finance. (2025). AI Shopping Is Here but Brands and Retailers Are Still on the Sidelines.

3. NIQ. (2025). Smarter Shopping: AI’s Role in Evolving Product Content Expectations.

4. IEEE Xplore. (2025). Generative AI-Driven Personalized Ad Content Generation Framework for E-Commerce Platforms.

5. AlixPartners. (2026). AI shopping agents: Customer gains, retailer gains, and what retailers stand to lose.

6. PayPal. (2026). What Merchants Need to Know About Agentic Commerce.

7. Digital Commerce 360. (2025). Ecommerce faces a structural reckoning in 2026.

8. Algolia. (2026). AI shopping assistants: A practical guide for ecommerce brands.