From recommendation to delegation: How AI is reshaping the consumer experience in digital markets

Agentic Commerce: Redefining the Consumer Experience Between Smart Efficiency and Trust Risks

Digital markets are undergoing a fundamental transformation in the consumer experience, as AI shifts from a “recommendation” tool to an “agent” capable of negotiating and purchasing on the user’s behalf. Estimates indicate that agentic commerce could reach $236 billion by 2034, with shopping agents projected to account for $190–385 billion of U.S. e-commerce sales by 2030—equivalent to 10–20% of market share. At the same time, experts warn that one-third of companies may damage the customer experience by deploying AI agents prematurely, putting consumer trust at risk. This article, based on recent academic research and industry reports, reviews AI mechanisms for enhancing the consumer experience—from hyper-personalization to autonomous purchasing agents—with a precise analysis of the tension between efficiency gains and transparency and trust risks.

First: From Personalization to Agency — The Evolving Role of AI in the Purchase Journey

1. Hyper-Personalization as the Foundation of an Enhanced Experience

AI’s journey in e-commerce began with recommendation systems using machine learning algorithms to segment consumer behavior and deliver personalized real-time experiences. An analytical study of millions of transactions from Amazon showed that AI-driven personalization produces a notable shift in spending patterns, with both price and quantity showing near-unity elastic response, confirming the economic and behavioral significance of these systems.

Platforms draw on a wide range of signals—reviews, prices, availability, delivery speed, return rates, and browsing history—to build an integrated picture of consumer preferences and deliver personalized recommendations. In this context, data consumers voluntarily share (“zero-party data”) stands out as a strategic asset, enabling brands to bypass cookie restrictions and build more personalized experiences.

2. Generative Experience — From Keyword Search to Intelligent Conversation

The shift from traditional keyword-based search to generative AI-powered search tools is a qualitative leap. Platforms no longer understand just the words a user types but now understand the intent behind the search. Tools like Amazon’s Rufus allow users to ask complex questions such as “Is this a good price?” or “Based on what you know about my daughter, recommend a birthday gift,” with the ability to track prices and purchase on the user’s behalf.

Amazon’s internal data indicate that more than 250 million users used Rufus over the year, with a 140% year-over-year increase in monthly users and a 210% increase in interactions. Strikingly, users who interact with Rufus during the shopping journey are more than 60% more likely to convert.

Second: Agentic Commerce — A New Model for Redefining the Purchase Journey

1. How Do Shopping Agents Work?

AI agents are autonomous software systems built on large language models capable of perceiving information, reasoning about problems, and acting across multiple steps without human guidance at every stage. These agents can compare thousands of options simultaneously, negotiate with multiple sellers in parallel, monitor markets over time, and make decisions based on user preferences rather than what the algorithm surfaces.

Researchers at MIT estimate that AI agents are poised to cause radical change in markets where human intermediaries already exist—such as real estate, insurance, and financial services—where decisions are high-stakes, the number of options is large, and information gaps between buyer and seller are wide. Researchers call this potential shift in transaction costs the “Coasean Singularity,” referencing economist Ronald Coase, who argued that transaction costs shape how firms and markets are organized.

2. The Battle for the Purchase Gateway — Amazon Between Expansion and Defense

Amazon has adopted an ambitious strategy to turn its interface into a unified purchase gateway beyond its own store. Through its “Buy for Me” feature, powered by Amazon Nova and Anthropic’s Claude, users can buy products from other retailers’ sites without leaving the Amazon app. The agent navigates to the third-party site, adds the product to the cart, enters the shipping and payment details stored with Amazon, and completes the purchase.

Products available through this feature have grown from 65,000 at launch to more than half a million today. The “Auto Buy” feature also lets users set a target price, while Rufus monitors products and executes deals automatically when the target price is reached, with early evidence of average 20% savings per purchase for Prime users.

Amazon’s withdrawal from Google Shopping ads in the U.S. (while continuing in Europe) reflects its confidence in Amazon as the primary gateway for product discovery, especially with Gartner projecting a 25% decline in global search traffic by 2026 as conversational solutions replace parts of traditional search behavior.

3. Platform-Owned Agents vs. User Agents

MIT researchers draw a critical distinction between two types of AI agents. First: the platform-owned agent, such as Amazon’s Rufus, which is convenient and ready to use but lacks transparency and creates strong incentives to prioritize platform interests, potentially leading to less beneficial outcomes for consumers, such as higher prices or fewer options.

Second: the “bring your own agent” (BYOA), a user-controlled, cross-platform portable agent designed to align with user preferences rather than seller goals. As researchers note, both models are technically feasible, and which one prevails will be determined by corporate and regulatory choices now—and these decisions are hard to reverse once entrenched.

Third: Consumer Gains vs. Trust Risks

1. Clear Gains: Saving Time, Effort, and Money

Field evidence provides a clear picture of agentic commerce’s benefits for consumers:

• Time 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 showed that users of Amazon’s “Auto Buy” feature saved an average of 20% per purchase.

• Expanded options: AI agents can compare thousands of options simultaneously, giving consumers access to a wider range of products and prices than was possible manually.

Visa research across three markets (U.S., Australia, New Zealand) showed that about one in three consumers expects to use AI shopping agents regularly. The main drivers were saving money, saving time, and discovering new products.

2. Trust and Transparency Risks — The Existential Challenge

Despite enormous benefits, experts warn that misapplying these technologies could significantly harm the customer experience. Forrester projects that one-third of companies deploying AI agents prematurely and in inappropriate contexts will damage the customer experience, weakening trust and jeopardizing customer acquisition and retention.

Visa research reveals three main consumer concerns:

1. Transparency: Nearly 85% of consumers consider it important to have clear visibility into how their data is collected and the ability to customize or delete their shopping and purchase information.

2. Loss of control: About half of consumers might stop using an AI agent if they lose control over how their data is used.

3. Accountability: Questions remain unanswered about liability for agent errors—such as misinterpreting user instructions or executing a wrong purchase.

A Meltwater and YouGov report found that 86% of consumers believe AI-generated content should be disclosed, and 32% trust brands that use AI-generated content less. The report also revealed that 51% of consumers remain unconvinced by AI, with misinformation concerns remaining high at 73%.

Fourth: The Path to Success — What Do Retailers and Brands Need?

1. Generative Experience Optimization (GEO)

The shift from Search Engine Optimization (SEO) to Generative Experience Optimization (GEO) is a fundamental transformation. Brands must 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.

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

• Visible trust signals: Such as review depth, ratings, delivery performance, and service reputation.

2. Trust and Transaction Infrastructure

As transactions move outside owned channels, payments and identity infrastructure become critical. Brands need to:

• Support agentic commerce protocols: Such as OpenAI’s ACP or Google’s UCP, to allow agents to transact on their behalf.

• Dynamic pricing and inventory systems: Because agents compare and negotiate in real time.

• Brand representation governance: Monitoring how external agents present the brand, its products, and prices.

3. Building Trust in the Age of Agents

Visa research indicates that trusted brands—such as PayPal, Amazon, and Visa—are most preferred by consumers to provide shopping agents, suggesting that pre-existing brand trust is a decisive factor in adopting these technologies. Experts recommend:

• Proactive transparency: Clearly disclosing how AI is used and how data is collected and used.

• Control and rollback mechanisms: Giving users options to control agent operations and reverse decisions.

• Security and accountability guarantees: Building systems that protect against fraud and misinterpretation, with clear liability in error cases.

Fifth: Conclusion and Future Outlook

AI is redefining the consumer experience in digital markets across four interconnected levels:

1. Personalization level: Moving from generic recommendations to hyper-personalization based on deep understanding of consumer preferences and behavior.

2. Search level: Moving from keyword search to intelligent conversation that understands intent and provides answers rather than links.

3. Agency level: The emergence of agents capable of negotiating and purchasing on the user’s behalf, with expectations they will account for up to 20% of e-commerce by 2030.

4. Trust level: Redefining trust in the digital age, where transparency, accountability, and control become prerequisites for adopting these technologies.

The message is clear: the future belongs not to those with the best recommendation algorithm alone, but to those who build an ecosystem combining smart efficiency with transparent trust. As the Visa report concluded: “The question is no longer whether consumers will use AI agents, but which player in the ecosystem they will trust and ultimately use.”

Sources and References

1. J.P. Morgan Payments. (2026). Agentic Commerce: The Future of AI-Powered Shopping.

2. ScienceDirect. (2025). AI-powered personalization in e-commerce: Governance, consumer behavior, and exploratory insights from big data analytics.

3. About Amazon. (2025). How Amazon is using generative and agentic AI to transform the shopping experience.

4. Forrester / Nasdaq. (2025). Forrester’s 2026 B2C Marketing, CX, & Digital Business Predictions.

5. MIT Initiative on the Digital Economy. (2026). How AI Agents for Commerce Could Shape Future Markets.

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

7. Visa. (2025). Earning Consumer Trust in Agentic Commerce.

8. Morgan Stanley. (2025). Agentic Shoppers to Boost E-Commerce.

9. Chicago Booth School of Business. (2025). How AI is Reshaping Personalized Marketing.

10. PayPal. (2026). AI and the New Storefront: What Merchants Need to Understand.

11. Meltwater / YouGov. (2026). Trust in the Age of Generative AI.