As an IT professional deeply immersed in the crossroads of artificial intelligence and commerce, Dominic Jainy offers a unique perspective on the technologies reshaping how we shop. With expertise in AI, machine learning, and blockchain, he has a keen eye for innovations that move beyond hype to deliver tangible business results. Today, we’re discussing a pioneering move by Debenhams Group to embed an “agentic AI” shopping experience directly within the PayPal app, a pilot program that could signal a major shift in online retail.
This conversation delves into how this integrated approach tackles the persistent problem of mobile cart abandonment by creating a seamless, contained shopping journey. We will explore the inner workings of the conversational AI that replaces traditional search, the crucial backend data infrastructure that ensures real-time accuracy, the strategic trade-offs of placing a storefront within a third-party ecosystem, and the internal upskilling required to manage this new era of commerce.
Mobile checkout abandonment is a persistent revenue leak for many retailers. How does keeping a shopper entirely within the PayPal ecosystem address this friction, and what early metrics suggest this compressed sales funnel is effective?
It’s a brilliant move to address a problem that has plagued e-commerce since the dawn of the smartphone. The core issue is friction. Every redirect, every time a user has to re-enter their shipping address or find their credit card, is a potential exit point. By containing the entire journey—from discovery to payment—within the PayPal app, you eliminate those leaky seams. The system automatically uses the saved account credentials for both delivery and payment, which means a purchase can be completed in a few taps within a single chat window. While it’s still in the pilot phase, the strategic rationale is solid, given that 16 percent of Debenhams Group’s sales already flow through PayPal. This isn’t about acquiring new users; it’s about better converting the ones you already have by meeting them where they are.
This new system bypasses traditional keyword search for a conversational agent. Could you walk us through how the agent uses a shopper’s profile to align recommendations with their budget and preferences, perhaps providing an example of the follow-up questions it asks?
This is where the ‘agentic’ nature of the AI really comes to life. Instead of you having to type “blue floral summer dress size 12,” you can simply say, “I need an outfit for a friend’s wedding next month.” The agent then accesses your profile data within the ecosystem—past purchases, price sensitivity, maybe even brands you’ve shown interest in. From there, it begins a dialogue. It might ask, “Is the wedding outdoors or indoors?” or “What’s your approximate budget for the dress?” to narrow the search. It’s moving from a search engine to a personal stylist, using conversational context to find the perfect item from brands like Karen Millen or PrettyLittleThing, which feels far more intuitive and personal than scrolling through pages of search results.
Given that agentic commerce relies heavily on real-time inventory and pricing data, how does the underlying AI infrastructure for forecasting support this automated experience?
You’ve hit on the most critical, and often invisible, part of the equation. An agent that recommends an out-of-stock item or displays an incorrect price is worse than useless—it actively damages customer trust. This is why Debenhams’ partnership with Peak AI is so significant. They are building a robust data lineage to support these automated interactions. This involves using AI to improve forecasting across stock levels, sales trends, and dynamic pricing. Essentially, they’re ensuring that the information the conversational agent relies on is a precise, real-time reflection of reality. This foundational work is what allows the agent to function accurately and avoid the kind of errors or “hallucinations” that would otherwise sink the entire customer experience.
This initiative places inventory discovery directly on a third-party platform where a large segment of your customers already operates. What are the strategic trade-offs of this approach versus driving traffic to proprietary storefronts, and how does it reshape your marketing strategy?
It’s a fascinating strategic pivot. The traditional model is all about driving traffic to your own website or app, where you control the entire environment. The trade-off here is giving up some of that control in exchange for being present where transactional intent is highest. They are positioning their inventory where the liquidity already exists, right inside the payment app. This fundamentally reshapes marketing. Instead of spending money on ads to pull customers to your site, you are integrating discovery directly into the point of settlement. The sales funnel is dramatically compressed, merging the top (inspiration) and the bottom (purchase) into a single, seamless workflow. It’s a bet that capturing high-intent traffic on a partner platform will be more efficient than trying to own every step of the journey yourself.
Alongside this customer-facing technology, you launched an internal AI Skills Academy. What specific skills are being prioritized to manage these new agentic workflows, and how do you measure the business impact of upskilling your teams in this way?
Launching advanced AI isn’t a “set it and forget it” project. The AI Skills Academy is crucial because it ensures the organization can actually manage and scale this new reality. The skills being prioritized go beyond basic digital literacy; we’re talking about training teams in applied AI. This means understanding how to monitor the conversational agent’s performance, interpret the data it generates, and fine-tune its behavior. It also involves managing the data pipelines that feed the AI and ensuring the backend systems for inventory and pricing are flawless. The business impact is measured by the successful operation of the agentic system. It’s about empowering your internal teams to be masters of the new technology, not just users of it, which is the only way to sustain innovation and avoid being wholly dependent on external vendors.
What is your forecast for agentic AI in commerce?
My forecast is that this pilot is the beginning of a fundamental transformation, much like the initial shift to mobile shopping was. We are moving away from the era of search bars and grids of products. The future of commerce is conversational. Shopping will feel less like a transaction and more like a dialogue with a trusted, incredibly knowledgeable personal assistant. This agentic layer will be embedded not just in payment apps, but potentially in a variety of platforms we use daily, from messaging apps to AI assistants like Microsoft Copilot. For retailers, the challenge will be to ensure their data and inventory are structured to plug into these conversational ecosystems, and for customers, it will mean a far more intuitive, personalized, and seamless way to shop.