Powering Agentic commerce and value added services through the unified customer ID (UCI)

A strategic outline for PSPs, Google, Loyyo partners.

1. Executive summary

The retail and digital commerce landscape is about to change. We are moving from human-driven browsing to agentic commerce (Adyen, 2026). In the near future, AI agents will execute commerce, discover products, and manage lifecycle purchases for consumers.

However, these AI agents currently have a massive contextual blindspot. While they can scrape the web for product data, they lack the actual historical purchase context of the individual consumer. Traditional customer data platforms (CDPs) attempt to fill this gap using probabilistic data, guessing user intent based on loyalty, ad clicks, cookies, and fragmented browsing history.

This paper introduces the framework for the unified customer ID (UCI). By bridging the gap between physical payment rails (PSPs), online purchases, and mobile wallets via Loyyo, you transform anonymous "ghost" transactions into clear, payment-linked profiles. The UCI provides the exact historical purchase context that AI agents require to execute true hyper-personalization, making probabilistic marketing obsolete and building the foundational data infrastructure for the agentic era.

2. The AI context crisis and the ghost transaction

Despite the digitization of commerce, a huge percentage of physical retail transactions remain ghosts. A customer walks in, taps their phone or card, completes a transaction, and walks out. The payment rail processes the transaction securely, but to the merchant's ecosystem, the customer remains completely anonymous. As industry leaders note, failing to identify these shoppers leaves substantial revenue on the table. Companies that excel at personalization generate 40 percent more revenue from those activities than average players (McKinsey & Company, 2021). Furthermore, omnichannel shoppers (those whose physical and digital footprints are connected) spend consistently more per transaction than single-channel buyers (PSPs, 2024).

The failure of the traditional CDP in the AI era To combat this anonymity, brands invest heavily in CDPs, attempting to stitch together a profile based on top-of-funnel tracking. Did they click an Instagram ad? Did they open an email? Did they browse a category page? Did they earn points or reach a membership tier? Did they earn points or reach a membership tier?

This probabilistic approach is flawed, especially as third-party cookie depreciation forces a shift to first-party data. It relies on guessing intent rather than acknowledging real actions. If an AI agent is tasked with acting on behalf of a user (for instance, restocking household supplies or recommending a specific accessory), it cannot rely on guesses. If the agent does not know what the user actually bought in the physical world, its digital recommendations will fail.

3. Bridging the gap: Creating the unified customer ID (UCI)

The solution is not more tracking pixels. It is accurate data capture at the point of sale, linked permanently to the user's primary digital identity device (their mobile wallet) and their payment. This perfectly aligns with PSPs's push for true omnichannel profiles that link point-of-sale and e-commerce into a single view.

The multi-layered identity resolution (the handshake)

1. The ghost tracking (Multi-token resolution): When a user taps their card or phone, multiple identifiers are generated. PSPs provide card tokens and additional shopper references. Simultaneously, card schemes (Visa, Mastercard) provide network tokens like the PAR (Payment Account Reference). Loyyo ingests these secure tokens to reliably track anonymous "ghost" behavior over time across different channels.
2. The customer enrollment: The PSPs terminal asks a customer to enroll, the cashier enters details into the POS, or the user adds a digital pass to their mobile wallet. Crucially, the wallet is not merely a local app on the phone; it is deeply tied to the user's primary authenticated account (their mobile OS ecosystem and browser).
3. The resolution: The moment the pass is provisioned, Loyyo retroactively links the persistent payment tokens (PAR, shopper_reference, PAN tokens) to the newly captured wallet device identifiers, member IDs, and the associated account email.

The Magic Loop: Activating the ecosystem The true power of the UCI is realized after this initial handshake. The wallet does not need to be scanned again to trigger personalization. The next time the user simply taps their bank card at the terminal, Loyyo instantly recognizes the PAR or card token and identifies the UCI. Because the UCI is linked to their Google Account via the wallet, the physical card tap instantly triggers the digital ecosystem. The merchant can push a highly personalized reward to their Google Wallet, or update their context so that when they open Chrome or talk to Gemini later that day, the AI already knows they just left the physical store.

The result is a flawless, payment-linked historical record. PSPs provide the secure tokenization and transaction data, but they cannot build the full picture alone. Loyyo ingests this raw payment data and enriches it by pulling in detailed line-item data (SKUs) from the merchant's point-of-sale (POS) systems, Order Management Systems (OMS), e-commerce platforms, and customer PII (names, emails, physical addresses). By mapping the scheme PAR, PSPs's shopper_reference, member IDs, and wallet identifiers to rich line-item data and real-time live data directly from schemes like Mastercard and Visa, Loyyo transforms this massive aggregated pool into a highly actionable dataset.

4. The 5 pillars of agentic commerce (powered by Loyyo)

With the UCI established within a universal commerce protocol, merchants can transition away from legacy marketing flows and fully embrace AI-driven commerce. As highlighted at the Google Wallet Merchant Connect EMEA Summit, agentic commerce fundamentally closes the gap between discovery and decision (Google, 2026). The Loyyo data infrastructure unlocks five distinct capabilities to achieve this.

Autonomous automation (The CDP killer) Marketing teams currently build complex, manual flowcharts. In the agentic era, agents monitor the UCI in real-time. Instead of a flowchart, the AI agent has a simple objective: maximize customer lifetime value (CLV). The AI sees a customer bought a high-end espresso machine via Loyyo's terminal data. Exactly 32 days later (the average time a bag of beans lasts), the agent autonomously generates a wallet pass notification: "Need a refill? Tap here to order your usual beans with 15% off." No human marketer built that flow. The AI did it based purely on the Loyyo purchase timestamp.

Behavioral micro-segmentation and RFM metrics This shifts marketing from demographic guessing to specific cohorts based on recency, frequency, and monetary (RFM) value. Loyyo unlocks the "ghost-to-known" history, meaning the moment a user joins the loyalty program, their entire past purchase history becomes visible. The agent can instantly segment users who bought winter gear in November 2024, but have never purchased a spring jacket, and use a mobile wallet. You aren't guessing what they like. You know exactly what they bought and when.

Predictive clustering This involves unsupervised AI clustering of real purchase receipts. An AI analyzing the Loyyo data might discover a cluster you never knew existed. For example, customers who buy running shoes on Saturday mornings are 80% more likely to buy premium recovery supplements on Monday afternoons. The agent then uses the customer's wallet to target that specific cluster on Mondays at 2:00 PM automatically.

Conversational insights Dashboards showing total sales are replaced by conversational queries. A merchant can simply ask their dashboard what gateway product converts unknown shoppers to loyalists the fastest. Because the UCI connects the ghost transactions to the known wallet, the AI can instantly tell you exactly which products serve as the best gateway items.

Intelligent agentic workflows Agents operating on a robust data foundation, such as a unified profile, can dramatically lower the barriers for customers, answering complex queries, retrieving transaction history, and facilitating direct checkout interactions across all channels. Generative AI is set to revolutionize retail marketing with the ability to automatically generate personalized content in real time, increasing loyalty and share of wallet while reducing costs (Commercetools & Google Cloud, 2024).

Agentic search reranking and dynamic pricing Search engines and AI chatbots talk to the Loyyo UCI before loading results or responding to inquiries. Because the UCI contains the user's membership tier (e.g., Gold or Silver) and their average purchase value (APV), the results are highly personalized. If a customer who consistently buys high-ticket items searches for "running gear", the search engine dynamically ranks premium, high-value products at the top. Furthermore, the pricing displayed on the website or within the chat can be dynamically adjusted to reflect their specific loyalty tier, offering exclusive discounts instantly. As noted in recent retail AI research, dynamic pricing optimization allows retailers to adjust pricing in real-time in response to demand and external events, while predictive AI crafts timely recommendations based on specific shopping histories to enhance engagement and conversion (Commercetools & Google Cloud, 2024).

5. Ecosystem synergy: Brains and brawn

The agentic era requires a deeply integrated tech stack. Payment processors like PSPs are critical, but they lack the detailed POS line-item visibility and cross-channel scheme data required to build a true unified customer profile on their own. That is where Loyyo provides the missing link. The UCI relies on three core components working in harmony.

• The payment rail (PSPs): The source of the raw transaction and the initial, secure payment network token.
• The brain (Loyyo): The data engine that builds the UCI. It aggregates PSPs data with POS or OMS line-items, e-commerce histories, and real-time live data from Mastercard and Visa. It captures the ghost transaction, holds the historical purchase data, and provides the actual context to the AI agent to close the gap between discovery and decision.
• The delivery (Loyyo): The unified delivery and engagement layer. It translates the AI agent's personalized offer into whichever surface the customer is on at that moment: a wallet pass on the lock screen, a push in the merchant's loyalty app, a dynamically priced product card in the webshop, a reranked answer inside an AI agent chat, or a tier-aware recommendation surfaced via Google or another search engine. The mobile wallet is one of several identifiers Loyyo uses to stitch the profile together; it is not the only delivery channel. Because every connected surface reads from the same UCI, points, status, offers, and recommendations stay consistent across every channel in real-time.

6. Conclusion and the future of retail infrastructure

The brands that win the next decade of retail will not be those with the largest advertising budgets. They will be those with the cleanest, most accurate data pipelines.

If you are not capturing zero-party and first-party data through payments and wallets now, your brand will be invisible to AI agents tomorrow. Adopting the unified customer ID framework through Loyyo ensures that as commerce becomes increasingly autonomous, your data infrastructure is already built to support it.

References

Adyen. (2024). The Adyen Retail Report: Powering the next era of unified commerce. Retrieved from https://www.adyen.com/knowledge-hub/guides/retail-report

Adyen. (2025). Agentic Commerce Has an Infrastructure Problem. Retrieved from https://www.adyen.com/landing/agentic-commerce-has-an-infrastructure-problem

Adyen. (2026). Agentic Commerce. Retrieved from https://www.adyen.com/knowledge-hub/agentic-commerce

Commercetools & Google Cloud. (2024). Ethical AI in Retail. Commercetools Resources.

Google. (2026, March). Google Wallet Merchant Connect EMEA Summit. Warsaw, Poland.

McKinsey & Company. (2021, November 12). The value of getting personalization right (or wrong) is multiplying. McKinsey Insights.