WTF is agentic commerce? | How brands can make loyalty and promotions agent-ready

This marks a fundamental shift in how shoppers discover, evaluate and purchase products. By 2030, agentic AI is projected to account for 15%–25% of total U.S. e-commerce, according to Bain & Company — an indication that businesses won’t truly be omnichannel unless they have incorporated agentic shopping experiences. 

In this new landscape, loyalty and incentive programs are becoming the defining factor in whether brands compete on differentiated value or on price. That’s because AI agents currently surface the easiest discounts to apply (or worse, discount codes that have been unofficially shared online), not the outcomes brands and retailers rely on like margin, retention or long-term customer value, which can be achieved through loyalty.  

For marketers, the key to competing and winning in agentic commerce is to become agent-ready now to deliver structured, agent-visible promotions and loyalty for the future. Laurens Van Wiele, Chief Product Officer at incentives platform Talon.One said the decisions brands make now will shape how they compete for years to come. 

“With any new technology, early adopters are usually rewarded for having already built a reputation and built the trust that they can handle the technology and are good at it,” Van Wiele said. “It’s important for brands to stay on top of the agentic commerce evolution and ensure that they don’t miss the first train.”

In this WTF explainer guide, Digiday and Talon.One break down how making loyalty and promotions programs agent-ready now gives brands the key to competing and winning in the era of agentic commerce.

Sure thing. Agentic commerce is a type of digital commerce in which autonomous AI systems act on behalf of customers to discover products, evaluate options, apply incentives and complete purchases. Essentially serving as virtual personal shoppers, these AI agents handle the shopping journey by researching products, comparing items and completing transactions with minimal human input — covering each step from discovery through checkout.

“Many people are already using tools like Gemini, ChatGPT or Claude for discovery,” Van Wiele said. “The next phases of agentic commerce will bring more aspects of the entire shopping flow directly into the AI assistant. It won’t just be about deciding which product you’re going to buy, but also completing the checkout within your assistant and seeing your order status.”

Agentic commerce differs from earlier forms of automated shopping in that AI agents don’t wait for human input to approve each step. They set a goal, evaluate options against structured criteria and execute on it. 

“Everything happens on one screen,” Van Wiele explained. “In a traditional model, shoppers browse websites or perhaps a top 10 ranking to decide what to buy and then place the order on a merchant’s website. Agentic commerce brings all of these aspects into a single window, which is the conversation with your AI assistant, like ChatGPT.”

Those are great questions. To really understand the ins and outs of agentic commerce, here are some key terms for brands and retailers.

• Agentic Commerce Protocol (ACP) is an open protocol introduced by OpenAI in September 2025 to enable agent-led purchasing flows, including instant checkout, through ChatGPT. It focuses on execution rather than business strategy.
  
• AI agent is an AI system that can set goals, reason over constraints and take actions autonomously. In commerce, agents manage shopping tasks such as comparison, checkout and post-purchase follow-up.
  
• Identity linking is a secure mechanism that allows an AI agent to associate a user with their merchant account. Identity linking allows the application of personalized promotions and loyalty rewards during agent-led shopping.
  
• Universal Commerce Protocol (UCP) is an open protocol introduced by Google in January 2026 to enable agent-led purchasing flows, including instant checkout through Gemini. It focuses on execution rather than business strategy.
  
• Unified Incentives Protocol (UIP) is a set of standards introduced by Talon.One in January 2026 that defines how to expose loyalty and promotions in a unified, machine-readable way across agent-led shopping journeys. UIP makes incentives discoverable, explainable and actionable for AI agents.

Agent-led shopping is still in its infancy, but up to $385 billion in U.S. e-commerce spend could be influenced by AI agents by 2030, according to Morgan Stanley. That’s a lot of ad spend. And it means purchase decisions will increasingly be made by AI agents that don’t read a brand’s homepage in the same way humans can — they can’t feel the brand’s story or respond to emotional cues. 

To prepare for this shift to agent-led shopping journeys, brands must ensure their promotions and loyalty are centralized and visible to AI agents in real time, with clear rules and customer identity attached, Van Wiele said.

“That’s the foundation that matters for everyone, ensuring your data is ready to be consumed,” Van Wiele said. “Agentic commerce is still being shaped, and there is still uncertainty around what the intricacies of agentic commerce experiences will look like once things standardize. But one thing is clear: a brand’s data has to be ready to be consumed.” 

When those foundations are in place, connecting to agentic commerce platforms becomes a straightforward integration, not a replatforming effort, Van Wiele added. 

What matters now is readiness.

No reinvention needed, just some technical refinements. Traditionally, brands have built their loyalty strategies to influence shoppers at the point of purchase, according to Talon.One CEO Christoph Gerber. These loyalty programs introduce differentiated value for members, such as exclusive access to products and events, customer recognition and long-term rewards. 

But traditional loyalty mechanics, including point balances, tier benefits and member-exclusive offers, are currently invisible to AI agents because they aren’t exposed as structured, machine-readable data in the same way as discounts.

Instead, the agents extract structured data and make decisions based on criteria like price and availability. They surface all discount offers — from leaked codes to expired campaigns to incentives that were never intended for broad use — and often steer consumers toward the lowest available price. 

Losing control of discounts is becoming a significant risk for retailers in agentic commerce, according to Gerber. “A big concern within agentic commerce is going to be fraud,” he said. “Coupon fraud is already a problem with people double-dipping loyalty bonuses and generic promotions, and AI will only make it easier to do that.”

Without a single system to govern which discounts AI agents surface and when, margins can erode quickly.  But if incentives are centralized and controlled, there will be fewer chances for abuse of discounts or promotions. When properly surfaced, membership data can enable AI agents to weigh a consumer’s brand preferences and perks alongside discounts, rather than defaulting to savings based on price alone, Gerber explained. 

“When AI agents are crawling the internet, they’re anonymous website visitors, and they only see what is online. But loyalty program members may have different rewards,” Gerber said. “An AI agent won’t know that unless it can report to the merchant that this member is looking for an item and request all the offers available for them. Then it becomes much more personalized.”

Agentic commerce favors brands that design their loyalty and promotions with these technical aspects in mind, according to Gerber. 

“Success in the AI world is that your data is ready and well structured, because an AI agent will make better suggestions if it can read information better,” he said. “That applies to product data, but, at the end of the day, consumers care about the quality of service — flexible return policies; fast shipping; loyalty program perks. These things are not visible to AI agents in a standardized way. That needs to change, and that’s what we’re working on.”

Google’s UCP enables AI agents, commerce platforms and merchant systems to work together from search to purchase. It covers core capabilities like product discovery, identity linking, checkout and order execution by defining how those shopping activities happen between AI platforms and merchants’ backend operations.

“UCP is a standardized way for merchants to expose data and communicate with Gemini, and platforms other than Gemini,” Van Wiele said. “The UCP defines data structures in terms of, this is what a product looks like and this is where you enter a coupon. It’s standardizing the communication between an AI platform and merchants’ different backend systems to ensure that data can flow bi-directionally.”

Brands’ loyalty programs are surfaced in UCP through identity linking, which allows an AI agent to securely associate a user with their merchant account, effectively enabling the agent to sign in to a brand’s membership program and benefits on the customer’s behalf.

“Identity linking is the key to making agentic commerce personalized,” Van Wiele said. “The status quo is searching for products and getting generic results. With identity linking, the AI assistant knows who that person is and can ask a brand for offers specifically for that person. Or, the assistant can tell the brand that this person is going to make a purchase and then see, for example, that the consumer is a gold-tier member. The consumer receives a more personalized experience from within the AI assistant window.”

Promotions in UCP are implemented as an extension of the checkout capability. When promotions are enabled, merchants can expose their discount data to AI agents so discounts can be discovered, displayed and applied directly within agent-led shopping experiences, rather than surfaced through exhaustive web crawling. 

“The merchant system has to explain to the AI agent which SKUs they have on sale. In the other direction, the AI agent needs to tell the merchant: This person is going to buy these products, they’re going to apply this coupon and this is their delivery address, etc.,” Van Wiele said. “It’s essentially trying to standardize the data structure for that communication between the merchant and the AI agent.”

Protocols like UCP define the foundation for how agents interact with merchants, but they will never be able to represent the full depth of each capability.

“UCP will support loyalty soon, but the reality is that no matter how much base functionality is added, there will always be aspects that are unique to individual platforms,” Van Wiele said. “Our clients communicate very specific things, like the amount of points needed to reach the next tier or a buy three, pay for two promotion. It’s not realistic to expect the UCP standard to cover all of those things from day one. So, Talon.One developed an extension.”

Talon.One’s UIP surfaces loyalty and promotions mechanisms in a unified way across AI agent-based shopping journeys. It gives AI agents access to the same incentives customers see in any other channel, such as personalized promotions, discounts, bundles, loyalty points, perks and program benefits. The AI agents are able to understand all incentivization methods in a standardized, machine-readable format.

“Our mission is to enable our clients to be successful in agentic commerce, to ensure that all of the experiences and incentives they provide to shoppers on their websites or apps are also available in the new agentic shopping channels,” Gerber said. 

The first standard Talon.One released is a loyalty extension that ensures that when people shop via AI platforms like Gemini, they benefit from a brand’s complete loyalty offerings. That includes giving AI agents visibility into a customer’s point balance and tier status; highlighting how many points a customer will earn or spend in a transaction; and supporting card-based loyalty programs that enable loyalty without identity linking.

“This loyalty extension is the first building block of UIP, with more to follow,” Gerber said. “Our ambition is that every Talon.One incentives mechanism — personalized promotions, enterprise loyalty management, offer management and execution — will be fully leveraged in agentic channels through UIP.” 

With those strong foundations in place, standards like UCP and UIP become agentic commerce accelerators, according to Gerber.

“At the end of the day, we don’t want UIP to just be UCP extensions,” he said. “We want it to be open for other protocols and standards that emerge that support a similar model of extendability. We want to be there and help our clients be successful there as well.”

Yes, absolutely. Agentic readiness isn’t about rushing to adopt new protocols. It starts with getting the fundamentals right — centralizing and governing incentives so they work consistently across channels.

Brands that take the following steps now to make their loyalty and promotions programs agent-ready will be best prepared to compete and win in the agentic commerce era.

1. Align teams early. Treat agentic readiness like any major digital transformation: align e-commerce, marketing and data teams; set priorities, including legal and privacy around identity and consent; and start with a focused pilot — then expand as learnings increase.
   
2. Establish a single system for managing incentives. Many companies still manage incentives across separate promotions platforms, loyalty program software, e-commerce campaigns and point-of-sale operations. Humans can navigate these gaps, but AI agents cannot. Centralizing decision-making enables consistent rules, tighter margin control and more predictable outcomes.
   
3. Make incentives performant and easily discoverable. When promotion and loyalty systems are slow or lack structured, machine-readable data, AI agents can’t surface offers. This forces brands to compete on price or risk being excluded. Treating discoverability as a requirement ensures incentives are fast, machine-readable and available at the moment of decision.
   
4. Differentiate loyalty beyond price. Brands that win in agent-led environments will build value beyond discounts — through tier status, access, recognition and experiences that agents can factor into decisions without relying solely on price. Brands should focus on benefits that add clear value and convert agent-driven purchases into long-term relationships.
   
5. Use loyalty perks to drive identity linking. In agentic commerce, brands rely on identity linking to stay connected to customers. UCP’s identity linking, for example, allows shoppers to link membership accounts to AI agents, unlocking points and personalized offers. To encourage opt-in, brands should use loyalty perks to reward consumer identification and consent across channels.

Agentic commerce is still in the early stages of development and expansion. More work needs to be done before end-to-end shopping is fully enabled on AI platforms and widely adopted by brands and retailers, but the technology is advancing rapidly. 

OpenAI, for example, released its ACP in fall 2025 to enable agent-led shopping, including instant checkout, through ChatGPT. The feature launched with Etsy sellers with an intended rollout to Shopify’s merchants in early 2026. However, OpenAI recently pulled back on that approach and, while Shopify merchants’ products still appear inside ChatGPT conversations, buyers typically still complete purchases on the merchant’s own online storefront.

“There have been some checkout prototypes — Etsy within ChatGPT and Shopify,” Van Wiele said. “Everything is pointing in the direction that by the end of the year, there will be multiple brands whose products you can purchase directly from at least one of these AI solutions.” 

Overall, agentic commerce’s nascency makes it hard to truly predict what comes next, Van Wiele cautioned. 

“We’re talking about an ever-evolving world, with ever-evolving standards,” he said. “A really standardized process, which is what Google’s UCP is trying to achieve and widespread adoption, will come next year. It will be an interesting journey to see how it evolves because so much has changed in just the last year.”

---

About Talon.One

Talon.One is the most powerful incentives engine that unifies loyalty, promotions and gamification into one holistic platform. Backed by enterprise-grade security and scalability, Talon.One empowers companies to build personalized, profitable promotions and loyalty programs using any data. Today, over 300 of the world’s most-loved brands, including Adidas, Sephora and Carlsberg, work with Talon.One to drive deeper engagement and lasting loyalty with their customers. To learn more, visit www.talon.one