Agentic Commerce: Redefining the Consumer Journey, Liability, and Loyalty

The rise of agentic commerce, and a world of artificial intelligence (AI) agents that research, recommend, and even purchase on behalf of consumers, is set to transform the payments ecosystem. For merchants, issuers, and consumers, this new paradigm raises pressing questions:

Who is liable when agents make mistakes?
How does authentication work when the “buyer” is an agent, easily mistaken as a bot?
What happens to loyalty programs and stored payment credentials?

Let’s explore agentic commerce through the consumer journey, from product discovery to returning customer loyalty, and unpack what it means for liability, authentication, and the future of the merchant-consumer relationship.

Pre-Purchase: Consumer Intent and Product Discovery

The first stage of the commerce journey is intent: a consumer knows (or believes they know) what they want. Traditionally, this intent is expressed through search queries, browsing merchant websites, or in-store visits. In agentic commerce, the consumer’s intent is filtered, interpreted, and acted on by an AI agent.

Liability in Physical Goods Transactions

Imagine a consumer asks their AI agent to order white sneakers. The agent mistakenly orders green sneakers. Upon delivery, the consumer wants to return them.

Today’s framework: The consumer is responsible unless the merchant erred in fulfillment. Return shipping policies often recoup the cost from the buyer.
Tomorrow’s framework: With agentic commerce, liability likely remains with the consumer as the agent acts as their authorized proxy. But consumers may be unhappy with this outcome: “If an agent made a mistake, why blame me?”. However, this problem may become an opportunity: AI platforms could introduce purchase liability coverage as a competitive differentiator (“if our agent makes a mistake, we cover the return”).

This shifts liability into a product feature rather than a legal requirement, creating opportunities for differentiation across platforms.

Liability in Consumable Goods Transactions

Consumables raise a different issue. Picture a consumer ordering a cold brew for pickup. When they arrive, they insisted they wanted a latte. The cold brew is now spoiled and cannot be resold.

Today’s framework: The merchant absorbs the loss. Misorders in quick-service restaurants are common and typically fall on the operator.
Tomorrow’s framework: If an AI agent misinterpreted the request, liability attribution becomes murky. Merchants may push back, arguing that the consumer (through their agent) bears responsibility. Unless regulation dictates otherwise, consumers remain accountable for agent mistakes. In these cases, agents could also assume the purchase liability as per our previous example; however, this flow may become prohibitively expensive for the agent to manage.

VGS Insight

VGS distinguishes between two categories of agentic commerce:

Chat-based commerce: The consumer engages directly with a conversational interface, is shown a SKU, and clicks “buy.” This resembles a marketplace referral/upsell. Merchants view this as incremental revenue that would not otherwise exist, so concerns around loyalty and card-on-file (COF) are minimized.
True agent-based commerce: The consumer delegates with intent (e.g., “order my usual from Starbucks”). Here, liability becomes critical, since the purchase directly replaces an existing merchant-customer relationship.

Purchase: Authentication and Payment Flows

The next stage is payment initiation and payment processing, ensuring the buyer is who they claim to be, and is authorized to transact.

The Authentication Dilemma

When does authentication happen in agentic commerce? Options include:

At account set-up (for true agent-led transactions): The consumer authenticates when granting the agent ongoing purchasing authority.
At each purchase (human-in-the-loop transactions): The consumer confirms every order (biometric prompt, push notification). Per the term regularly used in the industry, this would be considered an authentication flow dedicated to human-in-the-loop transactions.
Hybrid approach: Small, recurring purchases flow seamlessly; larger or unusual purchases require reauthentication.

The industry has yet to decide on an authentication standard. In the meantime, we can expect a diverse set of options. The solution will hinge on consumer adoption and which flow feels most natural and secure on which agentic platform.

Merchant Trust and “Know Your Agent”

In agentic commerce, it is important to implement a "Know Your Agent" approach. Some merchants may require agents to identify themselves, and without this verification, agents risk being misclassified as malicious bots. Ensuring that agents can be properly distinguished helps build trust, prevents unnecessary friction, and maintains smooth interactions between merchants and legitimate agents.

Merchants must trust that the consumer has authorized the agent and that the payments credentials are valid. Today, tokenization and strong customer authentication (SCA) help, but agent-driven commerce adds new intermediaries.

VGS Insight

Tokens will likely serve as the connective tissue. By mapping consumer credentials to authentication and payment methods, tokens can assure merchants that the transaction is legitimate even when an agent is in the middle. Until standards emerge, tokenization remains the safest bridge for merchants.

Post-Purchase: Disputes, Liability, and Upsells

Once the purchase is complete, the next challenge is error handling and dispute resolution.

Disputes and Liability

If a consumer disputes a charge, such as claiming they didn’t order the product, who is responsible?

Today: If the merchant fulfilled correctly, they usually win disputes.
Tomorrow: If an agent made the error, the same result likely applies: the merchant prevails, since the consumer authorized the agent. By delegating authority, the consumer essentially “signs away” certain rights.

Some AI agents may differentiate themselves by absorbing liability for errors or “hallucinations.” This could become a premium feature (“guaranteed accuracy or we refund you”).

Upsells and Dynamic Offers

Post-purchase engagement is a key revenue driver for merchants. Loyalty reminders, cross-sell prompts, and ‘people also bought’ campaigns all hinge on direct access to the consumer.

In agentic commerce, AI platforms may control these touchpoints. Upsell opportunities could be curated by the agent, not the merchant. Merchants may find their user experience and product marketing budgets competing with AI-driven product rankings.

Returning Customers: Loyalty, Card-on-File, and Merchant of Record

This stage highlights the deepest disruptions from agentic commerce.

Loyalty and Dynamic Experiences

Merchants collectively invest billions to build personalized customer experiences. Loyalty programs dynamically adjust offers based on historical purchase data, payment methods, and segmentation.

If AI platforms control discovery, loyalty shifts upstream. The agent, not the merchant, decides which merchants and products are displayed. Loyalty may migrate to the platform layer, with agents acting as gatekeepers.

VGS Insight

By linking loyalty IDs to payment credentials, the gap between a merchant’s existing loyalty systems and agentic payments can be bridged, keeping merchants and consumers aligned.

Card-on-File (COF)

Merchants prefer repeat customers to log in to their accounts and use stored payment credentials. This secures both loyalty integration and fraud reduction.

But how can an agent log in on behalf of the consumer? One solution is to bind COF credentials to the agent itself via tokenization. This allows the agent to act seamlessly with the consumer’s loyalty account.

Merchant of Record (MOR)

The MOR designation, which owns the transaction and assumes responsibility for chargebacks, taxes, and compliance, may diverge across platforms.

Some merchants will insist on being MOR to maintain compliance and control.
Others may allow AI platforms to act as MOR, effectively turning the transaction into a marketplace sale.

This could limit consumer choice: certain AI platforms may not integrate with merchants who require direct MOR control.

VGS Insight

Ultimately, AI platforms and the payments industry at large could adopt a hybrid approach. Ambitious AI platforms may attempt to cover the full commerce stack, from discovery to MOR. Others may stick to a marketplace model, passing MOR responsibility back to merchants. Either way, merchants will need to strategically decide where to draw the line between control and accessibility.

Looking Ahead: Standards, Trust, and Competitive Differentiation

The emergence of agentic commerce doesn’t erase the old questions; it reframes them. Liability, authentication, loyalty, and MOR have always mattered in commerce, but now they are being renegotiated with new players in the mix.

Key Considerations for Stakeholders

Consumers: Delegating commerce to agents creates convenience but reduces direct control. Expect platforms to differentiate by offering liability guarantees and custom features for unique user preferences.
Merchants: Trust and loyalty will be harder to maintain in a platform-first environment. Success depends on closing identity gaps (linking payment, authentication, and loyalty credentials) and adapting to new discovery models.
AI Platforms: The most ambitious will move beyond discovery and take on MOR responsibilities, competing directly with marketplaces and merchants alike.

VGS Perspective

At VGS, we believe the future of agentic commerce depends on securely binding authentication, payment, user identity, loyalty credentials, and more to ensure trust and flexibility. By tokenizing across the value chain, merchants and consumers can maintain control while embracing the convenience of agent-driven interactions.

Agentic commerce is not a question of if but how. Merchants who deliver frictionless convenience without sacrificing accountability will win.

Conclusion

Agentic commerce represents a seismic shift in the payments landscape. By reframing liability, authentication, loyalty, and merchant of record dynamics, it challenges both merchants and consumers to rethink established norms.

In pre-purchase discovery, agents redefine intent and responsibility.
During purchase flows, authentication standards remain unresolved.
In the post-purchase stage, liability and upsell opportunities evolve.
For returning customers, loyalty and COF face structural disruption.

The future is clear: AI-driven commerce will force new partnerships, new standards, and new expectations. The challenge, and the opportunity, for merchants is to adapt quickly while ensuring trust, security, and customer loyalty endure.