AI shopping is reshaping retail fulfillment as 45% of US consumers adopt it
Locus finds 45% of US consumers use AI to shop. Brand discovery and basket variance are reshaping forecasting, inventory placement, and returns.
Locus says its Q2 2026 US consumer survey found that 45% of consumers now use generative AI tools as a primary or secondary way to research or decide what to buy online. The company outlined the findings in its official survey post.
The more strategic point is not that AI is influencing shopping decisions. It is that AI is starting to change the shape of demand, and fulfillment systems are built to serve shapes they can predict.
Table of contents
Jump to each section:
- Why AI shopping changes the demand signal retailers plan against
- How brand discovery shifts inventory and fulfillment complexity
- Basket volatility is the new fulfillment constraint
- Returns operations are turning into a retention system
- What this means for marketers
Why AI shopping changes the demand signal retailers plan against
A common assumption is that AI shopping is mainly a top-of-funnel story: discovery improves, conversion changes, and marketing attribution gets messy. The operational reality is harsher. Fulfillment does not run on intent. It runs on forecastable demand patterns.
Locus’s data points to a threshold moment: when nearly half of shoppers use AI assistants, AI stops being an edge-case behavior and starts becoming a demand-shaping interface. Millennials are near 60% adoption, Gen Z at 50%, and within those groups meaningful shares treat AI as a primary tool (23% for Millennials, 17% for Gen Z).
Two strategic observations follow from that shift:
- When discovery moves into AI assistants, demand stops reflecting your storefront. It starts reflecting the assistant’s recommendations, which can re-rank brands and products faster than merchandising cycles.
- Forecast error is no longer just “bad modeling.” It is a structural outcome of a new demand interface.
That distinction matters because most retail planning stacks were tuned for historical continuity: last quarter looks like this quarter with manageable variance. AI-driven shopping introduces variance that is not random noise. It is behaviorally induced.
ContentGrip
How brand discovery shifts inventory and fulfillment complexity
One survey result should get operations and marketing leaders in the same room: 39% of consumers using AI to shop say they are more likely to try new brands or products they would not have considered otherwise. Locus contrasts that with 18% in the general consumer base.
This is not just “brand switching.” It is the expansion of the consideration set at the moment of purchase. Fulfillment feels that as brand mix unpredictability: emerging SKUs spike, legacy SKUs soften, and inventory positioning built on stable brand allocations starts producing the wrong kind of efficiency.
Here is the deeper tension: retailers often treat assortment expansion as a deliberate strategy. AI makes it partially involuntary. If assistants consistently surface certain brands for certain query types, the demand “migration” can look like a step change rather than a gradual trend.
A third strategic observation that is easy to miss:
- AI does not just increase choice. It increases the speed at which choice translates into operational load.
The operational knock-ons Locus highlights are practical: SKU proliferation, stockouts on newly demanded brands, and excess inventory on legacy brands when the underlying mix shifts faster than planning refresh cycles.
Basket volatility is the new fulfillment constraint
Locus’s survey shows what looks like a contradiction but is actually the point. Among AI shoppers:
- 37% are more likely to purchase more items in a single order (vs 17% general population)
- 34% feel more confident purchasing fewer items (vs 11% general population)
The implication is not that baskets are trending larger or smaller. It is that basket size distribution is widening. You get more orders at both tails.
Fulfillment systems are usually optimized around averages: average pick path, average cartonization, average weight, average last-mile capacity. Volatility is what breaks averages.
One-sentence reality check: variance is the cost center.
Locus links this to concrete operational stress: packaging mismatches, different picking dynamics, more multi-node sourcing when baskets span warehouses, and less predictable last-mile vehicle utilization. When larger baskets also include more diverse brands, complexity compounds, not adds.
Returns operations are turning into a retention system
AI-driven brand exploration has a predictable mirror image: more returns. Locus frames this as a reverse-logistics consequence of the same behavior that expands brand discovery.
The survey findings add texture to what “returns pressure” actually means:
- Bracketing behavior is low among Boomers (3%) but higher among Millennials (14%) and Gen Z (20%)
- 54% of consumers prefer in-store drop-offs (61% Boomers, 49% Millennials, 53% Gen Z)
- 68% say fast refunds make them more likely to shop with a retailer again
The operational insight is also a marketing insight: refund speed is no longer a back-office KPI. Locus positions it as a loyalty lever.
A fourth strategic observation that connects the dots:
- In AI-shaped commerce, the return is not the end of the journey. It is a moment that decides whether the customer will outsource their next decision to the same retailer or to the assistant.
If reverse logistics is engineered around cost containment only, it risks becoming a churn engine precisely as AI increases experimentation with unfamiliar brands.
What this means for marketers
AI shopping adoption is often discussed like a channel shift. Treat it more like an interface shift that reorders demand, basket composition, and returns. Marketing teams do not own fulfillment, but they increasingly own the promises that fulfillment has to keep.
- Reframe “AI shopping” as a demand-quality issue, not just discoveryIf 45% of shoppers use AI assistants in purchase research, the marketing signal (what shoppers want) and the operational signal (what gets ordered) will drift faster. Planning around stable patterns becomes a strategic liability.
- Brand consideration is widening, which changes how loyalty should be measuredWith 39% of AI shoppers more likely to try new brands, loyalty becomes more conditional. Brands should expect higher churn risk unless post-purchase experience and returns reinforce trust.
- Basket variability should influence how you message shipping, bundling, and availabilityWhen both larger-basket and smaller-focused purchases rise, one-size-fits-all messaging about delivery speed or thresholds can misfire. The constraint is not “speed.” It is predictability across a wider range of order shapes.
- Treat refund speed as part of brand experience designLocus’s 68% “fast refunds” finding reframes refunds as retention. Marketers can use this to argue for investment because it directly supports repeat purchase behavior, not just operational neatness.
- In-store drop-off preference is an experience signal, not a logistics footnoteIf 54% prefer in-store returns, returns partnerships and store workflows shape perception. That is marketing territory as much as operations.
The deeper shift is that AI assistants are becoming a decision layer between consumer intent and retailer execution. As that layer strengthens, brands will compete not only on product and price, but on whether their operations can keep the assistant’s promise consistently.
Retailers that treat AI-driven shopping as merely “new traffic” will keep optimizing the storefront. Retailers that treat it as “new demand physics” will redesign the system behind the storefront.
