The Agents are Here

This week, an AI agent at Amazon interviewed job candidates at 2 AM, and another one answered Crocs shoppers without a human stepping in. The agents are already working in retail, and the only question left is whether your data is clean enough to trust them. Hi, I'm Clare, 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. Before we get into the stories, mark your calendar for November 17th, 2026. That's Retail Agent Con. The agents are here, and retail is still deciding what to do about them. Agents are already changing how customers shop and how retail work gets done. And they are moving faster than most organisations have adapted. This half-day programme looks at what agents are doing in retail right now, which results are real, and how people and agents learn to work together in service of the customer and the business. If this week's episode hits close to home, you'll want to be there. The link is in the description below. This is a busy travel season for the IHL Analyst team. Check the website under Connect to make a connection. They would enjoy it. We grabbed Greg for this week's Big Picture.

Claire, the agents have left the strategy deck and entered daily operations. In a single week, we saw AI agents working with shoppers, running on new commerce rails, and helping hire the workforce. But there's a catch, and it's the thread that runs through every story in this segment. Agents act on whatever data you feed them. Now give us the details.

Let's start with the shopper. Crocs replaced its scripted self-service support with Rivet, a conversational assistant built on Salesforce Agentforce. Rivet brings product discovery, customer support, and order tracking together in one conversation. Crocs expects more conversations resolved without a human, better search success, higher click-through rates, and bigger baskets. And it plans to extend Rivet to its Hey Dude brand and to international markets with multilingual support. Crocs is ahead of a curve the rest of retail is climbing. IHL research on mid-market retailers finds customer service chatbots are deployed at 31% of them, and 58% of those deployments produced positive results. Chatbots also show the lowest share of retailers with no plans at all. And they're on track to lead deployments within 2 years. In other words, if you don't have one of these in the works, you're in a shrinking minority. Next, the commerce rails. Google is expanding its Universal Commerce Protocol, an open standard for agent-driven shopping inside Merchant Center. Retailers now get cart transfers back to their own sites and checkout testing, with more analytics coming. Tapestry is already live with select Coach and Kate Spade products purchasable directly from Google Search, AI Mode, and the Gemini app, just in time for the holidays. Here's why that matters. Agentic Checkout only works when clean product and order data flows through open standards and APIs, which is the heart of composable commerce. And the payoff for getting AI right is real. IHL research shows retailers using AI and machine learning posted 16% higher sales growth and 134% higher profit growth in 2024 than retailers that were not. Third, the workforce. Amazon Web Services made Amazon Connect Talent generally available. It's an agentic hiring platform that builds interview plans, runs voice interviews 24/7, scores candidates and hands the recruiter transcripts and notes. The recruiter still makes the call, and hiring decisions move from weeks to days. That lands at an interesting moment. In IHL's latest Retail Transformation Study, 60% of retailers expect fewer employees per store within 2 years, and they rate AI as one partial contributor among several. Meanwhile, sales leaders show a 95% greater expectation of adding employees. So for the retailers who are growing, faster and better-matched recruiting becomes a real competitive edge. Now let's go from the screen to the shelf. Petbarn, which runs 267 stores across Australia, is automating planogram management with AI-enabled Crisp connected to Blue Yonder's category knowledge base. The system handles more than 240 updates per hour and cuts prescription diet planning from 3 weeks to 1 day. Planogram execution is where brand promises break. IHL's shelf research puts average planogram compliance at just 57%, and 39% of retailers can't even reach 50%. Automating mass SKU changes attacks the labor bottleneck behind those numbers, And it gives brand partners the shelf consistency they're paying for. And that brings us to the least glamorous story of the week, and maybe the most important one. Aldi UK completed its move to an SAP ERP system after 2.5 years, with its final Irish stores going live. This is the work that makes AI and agents usable. IHL research finds 44% of retailers cite availability of real-time inventory data as a top challenge, and 25% cite a lack of system integration. Aldi UK now has a unified ERP foundation across its estate to build on. Greg put it plainly this week: retailers now dedicate 15% of their IT budgets to AI, and AI spend is rising 27% a year. Yet 38% still list data cleaning, training, and storage as a top priority. The agents are ready to work. The question is whether the data underneath them is ready too. Fix the data first, then let the agents run. All right, let's pick up the pace with our lightning round. Here's what else happened this week. Macy's is moving an AI forecasting overlay for inventory replenishment from pilot to broader rollout, aiming to improve in-stock levels as part of its Bold New Chapter plan. Instacart launched software that gives grocers and consumer brands real-time shelf data, built on more than 10 million data points a day from shoppers scanning shelves while they fill orders. Sam's Club added predictive targeting to its Sam's Club Connect retail media network, using membership data and more than 400 intent signals to forecast what members are likely to buy next. Asda will install up to 700 in-post parcel lockers across its UK stores, so shoppers can send, collect, and return parcels during their weekly shop. JD Sports is using Algolia as the foundation for its agentic commerce strategy. Since 2024, the retailer reports a 22% lift in search revenue and 73% higher click-through on product listing pages. Kingfisher, owner of B&Q and Screwfix, says AI-driven personalisation generated roughly £100 million in sales, and the group plans natural language search at B&Q and agent-enabled shopping next. Walmart is expanding its Scintilla supplier data platform, with customisable boards and a version of its Agentic AI assistant, and plans to bring Scintilla to Sam's Club Amazon's Project Mercury aims to grow same-day delivery hubs from 85 to more than 1,000 by 2031, putting a hub within 10 miles of 80% of US Prime members. Loblaw is scaling EAI GLE's vision— AI gate automation— across its distribution yards to cut processing times and improve data accuracy at the gate. Waitrose extended its partnership with Trust Retail through 2031, rolling out real-time inventory technology to about 125 stores before Christmas and the rest of the estate by March 2027. WorkJam is expanding across every Catalyst Brands banner, including JCPenney, reaching about 70,000 associates, and adding an AI agent for HR and benefits questions. And in restaurants, brands like Portillo's, Qdoba, and Bobby's Burgers are treating self-order kiosks as revenue engines, with research from the industry group showing kiosks can lift average quick-service orders by 15% to 30%. So here's the question to take into your week. If you handed an AI agent the keys to your inventory, your customer service, or your hiring tomorrow, would you trust the data it's working from? For more than 40 additional stories from this week, plus free white papers packed with IHL research, head over to ihlservices.com. Thanks for listening to Retail Reality Check. We hope to see you back next week.