AI is working its way through the buyer’s journey. The 2025 holiday season saw a 693% jump in AI-driven traffic from the previous year, as customers are increasingly relying on AI-based chat interfaces to help them in the initial search and discovery process for all manner of things, from buying clothes to planning trips, and more, even reaching further and further into the buyer’s journey to complete checkout.
With a projected market size of USD $65.47 billion by 2033, agentic commerce is rapidly moving beyond initial product discovery and checkout, as AI agents are increasingly designed to manage the entire lifecycle of a transaction, meaning autonomous software will soon take over routine post-purchase tasks like WISMO (Where Is My Order) requests, order modifications, and return initiations. This shift in the customer experience requires a fundamental restructuring of post-purchase operations to accommodate algorithmic shoppers.
Agentic commerce has been accelerated by the introduction of a few protocols, and while there are many more than the ones mentioned in this article, let’s start there:
Universal Commerce Protocol, or UCP, was created by Google and Shopify as an open standard that lets AI agents do everything from discovery to purchase from retailers in a secure way. This successfully navigates the current challenges that differing APIs and other systems cause for agents that need to be able to negotiate and buy from any number of providers.
Agent Payments Protocol, or AP2, is another open framework contributed to from a number of organizations (but announced by Google) and focuses on the payments part of a transaction and it was built to handle authorization and security issues that could easily arise with a sea of consumer and retail agents running largely unsupervised.
The Agentic Impact on Customer Service
So what does this mean for brands that need to fundamentally rethink who the “customer” is at specific points in the journey and how interacting with a human customer via a proxy changes the customer experience? Let’s explore a few aspects of this.
Full-Lifecycle Agent Orchestration
The underlying infrastructure of agentic commerce is built to handle much more than simply placing items in a shopping cart. The Universal Commerce Protocol, for instance, acts as a canonical transaction envelope that defines and manages everything from initial discovery and checkout sessions to the entire post-purchase lifecycle across merchants and agents.
In practice, UCP uses targeted webhooks to communicate order-tracking and fulfillment updates directly to the customer’s AI agent. This allows the machine to autonomously monitor delivery statuses and process logistics without requiring the human consumer ever to check an email or a tracking page.
Dispute Management in the AI Era
When an AI agent acts autonomously on behalf of a user, managing disputes, returns, and refunds introduces a new layer of liability. Emerging standards like the Agent Payments Protocol are explicitly designed to clarify this liability and reduce disputes through cryptographic consent proofs and strict auditability.
To survive in this ecosystem, merchants must ensure that their order confirmations, digital receipts, and return policy links are perfectly mappable to the consent artifacts preserved by AP2. When disputes inevitably arise, having fast, programmatic alignment between a merchant’s records and AP2’s audit trail will be critical to lowering operational pain and preventing financial losses.
Transforming the Operating Model
As AI agents assume responsibility for routine post-purchase interactions, the human workforce within the CX and operations departments needs to evolve. Forward-thinking brands are already actively preparing for this shift in the human dimension of commerce, where customer service team members will transition away from manually performing tasks, such as processing straightforward returns, and toward supervising AI teams.
This transition from viewing AI as an assistant to an autonomous “colleague” will require robust new governance frameworks and change management strategies to operate these systems in regulated, customer-facing environments safely.
What Leaders Should Do
To adapt to this new reality, brands need to prioritize making their digital infrastructure agent-ready (adopting tools that use common protocols like UCP and AP2, for instance), while maintaining their ideally customer-friendly components as well. This means that leaders must treat their shipping SLAs, return terms, and product constraints as machine-readable data fields rather than human-readable prose to serve agentic service best.
Don’t get filtered out.
If an AI agent cannot seamlessly parse a return policy through structured feeds and APIs, it will filter out the merchant entirely.
While a determined human may decide to proactively contact customer support or even go to a brick-and-mortar store instead, an agent will likely pass over an unreadable option altogether.
Don’t skip a step.
Organizations that optimize for agents integrate their backend fulfillment and return APIs directly with agentic protocols like UCP to facilitate seamless, automated resolutions for all routine post-purchase requests.
Again, think of agents as a much less forgiving type of customer where success is binary.
Don’t forget the human.
As importantly, what doesn’t change is that the end customer is still a human. While their AI assistant and other agentic processes may increasingly be involved throughout the buying and post-purchase process, brands still need to ensure that the human customer has an optimal experience.
This can be a challenge when customers do not spend as much time within a brand’s ecosystem during the initial buying stages. Still, it means that organizations need to find better ways to quickly identify and serve customers and their agents, wherever they may be in the journey.
Summary
For an autonomous shopping agent, a frictionless post-purchase experience provides the trust signal that indicates all systems can proceed. Automating this critical part of the buyer journey and ensuring that backend systems speak the language of the emerging commercial protocols, brands can drastically lower their customer service overhead while guaranteeing that algorithms confidently choose their store for the next purchase. It’s a win-win, and it will set the company ahead in the competition for agentic customers in the months ahead.
