BACK

Groceryshop 2026: From Signal to Action: The New Agenda for Retail

Groceryshop, held in Las Vegas in September, brought together retailers, technology companies, startups, and research firms, offering another close look at the industry's agenda. As expected, artificial intelligence was present in almost every session. In previous years, the focus was more on what AI could do; this year, however, business results began to take center stage. Sessions discussed how to make the customer's shopping experience easier, how to accelerate the thousands of daily decisions that need to be made in the store, and how to turn data into actionable insights rather than just displaying it. Looking back at my notes, five topics stand out from my perspective.

1. Speed ​​is no longer just delivery speed.

   In retail, when we talk about speed, we've often discussed how many minutes it takes for an order to be delivered. At Groceryshop, however, a different concept of speed has emerged: reducing the time a customer spends shopping. Hy-Vee's example was striking. Today, it takes their customers about 40 minutes to complete an online purchase; the goal is to reduce this to five minutes. This is where the real opportunity lies for Agentic AI. Instead of searching for products individually, the customer will simply say, "I want to prepare a healthy meal for four tonight." The AI ​​will then prepare the shopping cart, taking into account the menu, preferences, budget, and available stock. Initial results are promising. DoorDash shared that customers using Ask DoorDash have 50% larger shopping carts. Kroger also reports seeing larger carts among customers using the AI ​​Shopping Assistant. So, AI isn't just speeding up shopping; it's also increasing cart size by reducing the burden of decision-making. Perhaps in the coming years, one of the key KPIs of online retail will be the time a customer spends completing a purchase.

2. A good AI experience requires strong operations.

   A message from Kroger's session was particularly important. You can offer a customer the best product recommendation in the world. But if the recommended product is out of stock or cannot be fulfilled during the ordering process, the entire experience is instantly ruined. Therefore, customer-side AI needs to work in conjunction with stock, store, product, and fulfillment information. This is a point sometimes overlooked in agentic commerce discussions. Building a very powerful AI assistant on the front end is not enough on its own. If there is no near real-time stock information, product data, pricing, promotions, and operational systems behind it, the AI's promise cannot be fulfilled. In short, the AI ​​experience is only as strong as the operations behind it.

3. We are moving from data to action.

   This was the area that interested me most, given the long-standing work we've done at OBASE. For years, we've been producing excellent reports and dashboards in retail. When a problem arises, the system shows it. Then, someone needs to look at the screen, understand the problem, make a decision, and take action. The new approach aims to shorten this chain. One of the standout initiatives presented a structure that continuously monitors activity within the business. This structure identifies sales, profit, or inventory opportunities, suggests what needs to be done, and directs the action to the right person; moreover, it tracks whether the outcome has been achieved. Another initiative focused on expiration date management. It digitizes product expiration dates and generates alerts only when intervention is needed, instead of requiring store employees to constantly check shelves. According to shared results, this approach reduced shelf inventory check time by approximately 75 percent. This provides an important clue about the direction store operations will take. The role of AI might be less about replacing the employee and more about telling them which store, which shelf, which product, and when intervention is needed. Indeed, according to Groceryshop's research, 85% of supermarket managers believe that roles working with AI will evolve into areas such as customer advisor, product specialist, or experience manager.

4. Optimizing price, promotions, and product range individually is becoming increasingly difficult.

   The "domino effect of decisions," as we call it, was also on Groceryshop's agenda: decisions in retail are becoming increasingly interconnected. When you change the price of a product, demand changes. When demand changes, inventory, replenishment, alternative products, promotions, and even customer communication are affected. Despite this, a significant portion of these decisions are still made separately by different teams and systems. One of the messages from the event was that, thanks to AI, it's becoming increasingly possible to optimize product portfolio, pricing, and promotion decisions together at the store and SKU levels. One of the most accurate approaches I've heard on this topic was: First, define the business problem and the profitability opportunity, then decide which AI model is needed. How many AI pilots we conduct or which model we use isn't the answer. The real question is what problem we solve in sales, margin, inventory, or customer experience.

5. GLP-1 may be more important for food retail than we thought.

Aside from AI, the topic that most caught my attention at the event was GLP-1, which has been on my radar for the past two years. According to PwC's June 2026 research, the percentage of households in the US with at least one person using GLP-1 has rapidly increased from 9% to 21%. Data shared by Groceryshop shows that households starting to use GLP-1 experience a 5.3% decrease in grocery spending in the first six months; this rate rises to 8.2% in high-income households. In contrast, only 17% of industry executives expect a 5-10% decrease in sales due to GLP-1 in the next five years. However, the decrease in spending among user households is already over 5%. This means that a measurable change in consumer behavior has already begun while its use is still becoming widespread; the industry's expectations are lagging behind. Moreover, this may not only mean consuming less food. Category balances may also shift in areas such as protein, fresh produce, healthy snacks, portion sizes, and functional nutrition. Therefore, GLP-1 should not be seen solely as a matter for the health sector. This could influence many retail decisions in the coming years, from product ranges to category areas.

What does this mean for Turkish retail?

None of these points are unique to the US. In Türkiye, where online shopping is rapidly growing, it's possible to measure how much time a customer spends completing their shopping cart and link this to a target. However, the warning in the second point is perhaps even more important for us. Before setting up an AI assistant on the front end, we need to ask how up-to-date and reliable the stock, product, and price data it will carry is. On the store side, for our retailers operating with high employee turnover and extensive store networks, systems that answer the question of "when should we intervene in which shelf" can be a significant efficiency leverage. The use of GLP-1 is increasingly being discussed in Türkiye. Retailers who start monitoring this signal in category data today will not be caught unprepared when the pace of change accelerates.

Common point: The distance between decision and action is shortening.

When I put these five points together, Groceryshop's main message is quite clear. The fundamental problem in retail is no longer a lack of data. The real issue is how quickly we can understand the incoming signal, how quickly we can translate it into the right decision, and how quickly we can implement that decision within the operation. On the customer side, AI is trying to reduce a 40-minute shopping trip to five minutes. It tells the store employee what action to take. In category management, it's starting to evaluate price, promotions, and product range together. At the same time, new signals like GLP-1 are emerging that can rapidly change consumer behavior. Therefore, the competition in the coming period will not be determined solely by having more data or stronger AI models. Retailers who can shorten the distance between the signal and the action the most will gain an advantage.