AI Agents Will Recommend Products Your Stores Can't Find

70% of retailers have inventory accuracy problems every week. Now, an AI agent is about to promise your customers products your store can't find. Hi, I am Claire, your host. Thanks for joining this week's Retail Reality Check, a podcast on the latest tech installs in retail and consumer product goods. We provide hard data so you can make smart decisions based on IHL Group research. To set the stage, here is this week's reality check. The shopper's starting point is shifting from the retailer's website to wherever the question gets asked. Keep that line in mind because every story in the deep dive points back to it. Before we get started, mark October 20th at 11 AM Eastern. IHL is hosting a webinar called Beyond the Shrink: The 2026 Global Retail Loss Study. For years, retailers have talked about loss prevention as if it only meant theft. Theft is only 25% of total retail loss. This is a worldwide study of what retailers are facing with theft, supply chain issues, organized retail crime, violence, and more. You will hear the benchmarks, where the industry is heading, and how the economy is changing the numbers. We will start with an economic update on the state of retail. You don't want to miss it. The registration link is in the description below. Here is this week's theme. The place a shopper starts is shifting from the retailer's website to wherever the question gets asked. Constructor launched Agentic Checkout, and it claims to be first to offer Stripe-powered checkout directly inside an on-site shopping agent. A shopper can discover a product, ask questions, get a recommendation, and pay without ever leaving the agent. Sephora and the Very Group are among the retailers Constructor serves. Here is the IHL angle. Among mid-market retailers, 83% name AI for personalized experiences as one of their most exciting technologies. Yet 71% cite integration complexity with existing systems as a barrier, and 75% rank ease of implementation as important or critical. Putting payment inside the agent bundles discovery and checkout, which lowers that burden. Next, Walmart joined OpenAI's new Virtual Try-On and Favorites in ChatGPT from day one. Shoppers can use a selfie, a screenshot, or a product image to see clothing on themselves, then save items before they buy. IHL's 2026 research shows why this matters. Personalization ranks first among retail technology priorities at 58% of retailers, a spot it has held in 7 of the last 9 years. Retailers now put an average of 15% of IT budgets into AI. Walmart's move shows where that money is landing: third-party AI surfaces. Gap is moving the same direction. Old Navy and Banana Republic are rolling out conversational product discovery. Gap is partnering with Daydream and Alta Daily for natural language search and avatar-based virtual styling. Its Google Cloud Foundation supports AI-native checkout through Google Search and Gemini. IHL research shows that retailers using AI and machine learning posted 134% higher profit growth than those that don't. The same report warns that weak performers rushing into generative AI without clean data face garbage in, garbage out. Gap rebuilt on unified Google Cloud data first, which is the right sequence. Now a reminder that the foundation matters as much as the front end. Rutter's, a 94-store convenience chain in the Mid-Atlantic, is deploying the full NCR Voyix Commerce Platform to unify convenience store, food service, and fuel workflows. IHL found that leaders are 92% more likely to run POS and mobile POS software on a modern architecture with microservices and a single codebase. And that microservices adopters report 83% higher profit growth. Rutgers fits the leader pattern. Last in the deep dive is a resort in the Bavarian Alps. Schloss Elmau cut reservations administration by 50% using Oracle's Opera Cloud and embedded AI bots that handle reporting, compliance, and routine communications. Staff see live guest profiles on their phones. IHL's 2026 study found that mobile devices for managers and associates correlate with 93% higher profit growth, second only to geolocation. Hospitality respondents also report the highest early adopter share, at 32%. The takeaway is the same across all 5 stories. Before you chase the next AI channel, confirm your product, price, and stock data can answer a shopper correctly on the first try. IHL's 2026 Transformation Study advises piloting in 5 to 10 stores and measuring results over 90 days before you scale. Now we move to our lightning round for a quick look at this week's news. Snowflake says retailers can improve forecasting, allocation, pricing, and personalization with AI, but only when governed data and consistent metrics are in place. Rinascente uses Snowflake to unify its customer data and speed up reporting. DoorDash launched text-to-order in the US, so users can search, build a cart, and check out over SMS. It also added one-tap grocery restocking, drone delivery with Chipotle and Popeyes, and a Dasher returns service. Albert Heijn, owned by Ahold Delhaize, is advancing agentic commerce with Stijn, an AI assistant in its app. It has grown from recipe help to weekly meal planning, basket prefill, and preference memory. Kroger worked with VidMob and MNMA Global on AI-driven predictive creative scoring across 1,900 ad assets. The approach forecast ecommerce conversion with 81% accuracy, and recommended creative delivered 4 times higher conversion rates. McDonald's next strategy commits $8.5 billion in partner support through 2036. It includes Google-built generative AI for drive-thru voice ordering, with a target of 250 basis points of restaurant-level efficiency gains. Albertsons is expanding its OpenAI partnership. A new Safeway experience in ChatGPT lets shoppers build grocery carts from recipes, photos, or lists and see their savings. Lidl is rolling out new self-checkouts across UK stores. They use overhead cameras and large screens that replay scanning footage. And the Lidl& Go self-scanning tool is expanding to 37 more stores. Cotswold Outdoor Group launched an in-store retail media network with Broadsign and IUF. It covers 149 displays in 23 stores and went live within 12 weeks. SEWA Logistics, Zalando, and Seriact launched automated fashion returns processing in Germany and Poland. Dual-arm robots identify, grasp, and sort unfamiliar items without article-specific training. Stussy picked RELEX solutions to unify demand forecasting, replenishment, and merchandise planning. Telia also chose Relex to replace spreadsheet-based planning across its retail stores and distribution centers. Winnow launched Winnow Foresight with Hilton, an app that builds daily prep plans for select-service hotels. European Hilton pilots reduced leftover food weight by up to 69% and overall food costs by 5%. Jysk rolled out Enacta's unified commerce platform across all 110 Norwegian stores. The launch follows Sweden and Finland, with the Netherlands and Belgium next. Amazon and the US Postal Service are piloting same-day delivery starting in West Virginia and Arizona, with South Carolina planned next. That wraps up the top news this week. So here is a question to take with you. When an AI agent promises your customer a product, can your store deliver it? Because the shopper will not blame the agent. They will blame you. It is all about your reputation. You can find the rest of this week's story list at ihlservices.com, with more than 40 additional stories and free whitepapers on the topics we covered today. Thanks for listening, and we hope to see you back next week.