Locus 2026 US Consumer Survey: Generative AI isn’t Just Changing How Consumers Shop, it’s Breaking the Demand Patterns US Retail Was Built On

The arrival of generative AI as a mainstream consumer shopping tool has produced one of the more significant behavioral shifts in US retail in the last decade. The Locus Q2 2026 US Consumer Survey found that 45% of US consumers now use AI tools like ChatGPT, Claude, and other AI assistants, as either a primary or secondary tool when researching or deciding what to buy online. The adoption isn’t evenly distributed. Nearly 60% of Millennials and 50% of Gen Z use AI to shop online, with 23% of Millennials and 17% of Gen Z treating AI as their primary research tool. Men adopt AI as a primary shopping tool at nearly 20% across generations; women adopt at 12%. The behavior is no longer emerging, it’s the operational reality retail fulfillment now serves.

Most analysis of AI shopping adoption focuses on the consumer behavior side — how AI changes product discovery, how it affects brand consideration, how it shapes purchase decisions. The analysis is real and useful. But the operational analysis — how AI-mediated shopping affects the demand patterns retail fulfillment infrastructure has to serve — has received less attention. The demand patterns AI shopping produces are operationally distinct from the demand patterns pre-AI online shopping produced. The distinction isn’t academic. It affects demand planning, inventory positioning, fulfillment network design, and last-mile delivery operations in ways that legacy fulfillment architecture wasn’t designed for.

The survey data shows three specific behavioral shifts that translate into operational pressure on fulfillment. Each shift is documented in the data; each has consequences for what supply chain operations actually have to handle.

Behavioral Shift 1: AI Surfaces Brands Consumers Wouldn’t Have Discovered

The first behavioral shift is the one with the most direct effect on demand predictability. As much as 39% of consumers using AI to shop are more likely to try new brands or products they wouldn’t have considered before. Compared to the general consumer base — where only 18% indicate the same propensity to brand exploration — AI-using shoppers are more than twice as likely to expand their brand consideration set in any given purchase.

The operational consequence is brand mix unpredictability. Demand planning systems that model brand-level demand based on historical patterns assume those patterns reflect a relatively stable brand consideration set among target consumers. When AI-mediated discovery surfaces brands consumers historically hadn’t considered, the brand mix shifts in directions historical patterns don’t predict. A category that’s been dominated by three or four established brands for years may suddenly see meaningful volume migrating to brands the demand planning system has limited historical data on. The migration isn’t gradual; AI-mediated discovery can produce step changes when an AI assistant consistently surfaces a particular brand for a particular query type.

Fulfillment operations feel this in specific ways. Inventory positioning calibrated to historical brand mix produces stockouts on emerging brands and excess on legacy brands as the AI-driven migration accelerates. SKU proliferation increases as retailers stock more brands to meet AI-surfaced demand. Forecasting accuracy degrades on category-level metrics because the brand mix underneath the category is shifting faster than the category-level forecast can capture.

Also Read: AI Route Optimization & Failed Deliveries: NA Analysis

The longer this behavioral pattern persists, the more it compounds. Each AI-mediated purchase produces consumer data the retailer’s systems may or may not capture; each surfaced brand becomes part of the consumer’s consideration set for future purchases; each successful AI-surfaced purchase reinforces the consumer’s reliance on AI for the next purchase decision. The operational effect isn’t a one-time adjustment; it’s a continuous shift in the demand patterns fulfillment serves.

Behavioral Shift 2: AI Encourages Larger Baskets in Some Segments

The second behavioral shift moves in the opposite direction from what intuition would suggest. Today, 37% of US consumers using AI to shop are more likely to purchase more items in a single order, compared to 17% of the general consumer base. AI assistants helping consumers research products often surface related items, complementary purchases, or category alternatives that consumers add to baskets they would otherwise have left smaller.

The operational consequence is basket composition variability. Fulfillment infrastructure calibrated to typical basket sizes, packaging optimized for average orders, picking processes calibrated to common item counts, last-mile capacity allocated to historical order weight distributions — produces operational inefficiency when basket composition shifts. Larger baskets require different packaging, often require different picking processes, occasionally require different fulfillment node sourcing (when items span warehouses), and produce different last-mile vehicle utilization patterns than smaller baskets.

The basket size shift compounds with the brand mix shift. Larger baskets with more diverse brand composition produce orders that pull from more fulfillment nodes, span more SKU categories, and require coordination across more supply chain partners than smaller baskets with concentrated brand mix. The fulfillment operation handling these orders faces materially more complexity than the operation handling traditional online retail baskets.

Today, 37% of US consumers using AI to shop are more likely to purchase more items in a single order, compared to 17% of the general consumer base. 

Behavioral Shift 3: AI Also Drives More Confident Smaller Purchases

The third behavioral shift introduces an apparent contradiction with the second. 34% of consumers using AI to shop feel more confident purchasing fewer items, compared to 11% of the general consumer base. AI assistants helping consumers make decisions sometimes lead to more focused purchases rather than larger ones — the assistant helps the consumer identify exactly what they need, and the consumer buys that rather than over-purchasing.

The contradiction is only apparent. Different consumer segments use AI differently for different purchase types. A consumer using AI to research a specific item may make a more focused purchase. A consumer using AI to discover a category may make a broader purchase. Both behaviors exist in the data simultaneously, producing demand patterns with higher variance — some orders larger, some orders more focused — rather than orders trending uniformly in either direction.

Also Read: Beyond CX: What North American Shippers Should Demand from Their Logistics Partners in 2026

The operational consequence is forecasting variance. Fulfillment infrastructure expecting predictable order size distribution faces wider distribution than historical patterns suggest. Operations calibrated to the mean order size produce inefficiency at both tails — over-packaged small orders and under-prepared large orders. The variance itself is the operational challenge, not any single behavioral direction.

Behavioral Shift 4: AI-Driven Brand Exploration Reshapes the Returns Surface

The three shifts above analyze the forward-logistics side of AI shopping. A fourth shift follows directly from the first — and lands in reverse logistics. When 39% of AI users try unfamiliar brands, the predictable consequence is higher return rates. Brand exploration and returns correlate; AI shopping reshapes both what enters the fulfillment network and what comes back.

The Locus Q2 2026 survey reveals how operationally consequential the reverse logistics dimension has become.

Returns volume is generationally fractured along the same lines as AI adoption. Only 3% of Boomers practice bracketing — ordering multiple variants intending to return some. The figure rises to 14% among Millennials and 20% among Gen Z. The generations adopting AI at the highest rates are also the generations producing return-heavy behaviors. As AI-driven brand exploration accelerates among younger shoppers, return volume scales with the same behavioral shifts.

Return channel preference is more concentrated than retailers assume. 54% of consumers prefer in-store drop-offs — across generations (61% of Boomers, 49% of Millennials, 53% of Gen Z) — dwarfing the 19% who choose whatever’s cheapest. Retailers building reverse logistics around home pickup or carrier-collection models are designing against consumer preference.

54% of consumers prefer in-store drop-offs across generations, making in-store partnerships infrastructure rather than convenience. 68% say fast refunds make them more likely to shop with that retailer again, moving refund speed from a finance workflow into a retention lever.

Refund speed is now a loyalty lever, not a back-office concern. 68% of consumers say a fast refund makes them more likely to shop with that retailer again. Refund speed sits in logistics and finance workflow, but the operational consequence sits in customer retention. Retailers absorbing AI-driven return volume without investing in refund speed convert their returns operation into a churn engine.

The operational implication: reverse logistics isn’t a separate problem from AI-driven forward-logistics shifts; it’s the same problem on the return side of the network

The Operational Implications

The behavioral shifts cluster into three operational implications for retail fulfillment leaders.

Demand planning has to model less predictable brand mix and basket composition. Demand planning systems built on historical brand-level demand patterns face structural limitations when AI-mediated discovery shifts the brand mix faster than historical patterns can capture. Operations leaders should examine whether demand planning operates against frequent-enough refresh cycles to catch AI-driven brand migration, whether the planning incorporates AI-mediated discovery signals (rather than relying purely on historical sales), and whether the planning models account for the variance higher AI shopping adoption produces rather than just the mean demand patterns.

Inventory positioning has to handle higher variance in actual demand. Inventory positioned based on historical brand allocations produces operational mismatch as AI-mediated discovery shifts consumer brand consideration. Operations leaders should examine whether inventory positioning accommodates faster brand mix evolution, whether SKU rationalization processes can keep pace with AI-driven brand introductions, and whether positioning logic handles the variance in basket composition AI shopping produces.

Also Read: $850B US Returns: AI Routing for Reverse Logistics 2026

Fulfillment infrastructure has to serve baskets with more diverse and less predictable composition. Picking processes, packaging allocation, multi-node sourcing logic, and last-mile capacity planning all face operational pressure when basket composition becomes less predictable. Operations leaders should examine whether fulfillment infrastructure is calibrated to the actual operational variance current consumer behavior produces — or to historical patterns that AI adoption has already shifted.

Reverse logistics has to absorb the predictable consequence of AI-mediated brand exploration. When AI shopping drives 39% of users to try unfamiliar brands, return rates rise — and the survey data points to the operational levers that determine outcomes. 54% of consumers prefer in-store drop-offs across generations, making in-store partnerships infrastructure rather than convenience. 68% say fast refunds make them more likely to shop with that retailer again, moving refund speed from a finance workflow into a retention lever. Operations leaders should examine whether reverse logistics is architected for the return volume AI-driven demand is now producing.

The cumulative diagnosis matters more than any single tactical adjustment. Retailers facing AI-driven demand pattern shifts need fulfillment architecture that adapts to evolving operational reality rather than fulfillment infrastructure that worked for pre-AI consumer behavior and requires tactical compensation for the gap between historical calibration and current reality.

The strategic question for US retail supply chain leaders is simple: with nearly half of consumers now using AI to shop online, is your fulfillment network built for AI-driven demand patterns—or for a shopping behavior that no longer reflects reality?

Learn more about how AI is transforming buying patterns and ultimately fulfillment, visit locus.sh