The vision of robots rolling through store aisles politely assisting customers might seem more compatible with reality than ever, particularly now that we can build bots capable of solving tasks organically. But the truth is that many shoppers, if not most, aren’t exactly sold on the idea of supplanting a warm-blooded staff with a mechanized alternative. Multiple experiments have revealed that consumers aren’t necessarily averse to using AI tools in-store, however. It’s just a matter of how the technology is deployed: namely, whether it promotes a personalized customer experience.

Some retailers, including Lowes and Kohl’s, have recently tested AI-equipped customer service bots. In Kohl’s case, the experiments failed before reaching the stage of customer interaction, but Lowe’s, on the other hand, developed and piloted a retail bot in several stores around San Francisco. The bots were found to be helpful when it came to completing routine jobs—like checking prices and scanning shelves for inventory purposes—but were less effective when fielding questions from customers. Lowes has since confirmed that while the AI-equipped bots served as useful tools for associates, and provided insight into store functionality, they won’t be reprising their role as customer assistants.

The problem with retail robots that mimic human behaviors or functions is the same factor that adds shock value to spirits, demons, and other creepy-yet-anthropomorphic creatures in horror movies. A being that closely resembles humans, but is still slightly off in either appearance or action falls into what researchers call “the Uncanny Valley,” or a realm of distorted realism that provokes visceral discomfort in human observers. A survey by RichRelevance seems to verify that the Uncanny Valley may explain, in large part, customers’ preference toward human retail staff. The survey found that nearly a third of consumers found the mere concept of in-store robots to be “creepy” instead of “cool.”

As a technology, AI offers undeniable benefits to retailers. Automated processes are quickly becoming invaluable in data analytics and insight generation. In fact, one of retailers’ top five tech goals is to harness AI’s potential to fuel data gathering and better inform decisions,  according to an analysis from Forrester Research.

While stocking stores with robotic armies of walking, talking AI may not be the ideal route, this technology has proven successful in cases where it deepens personal experience, rather than complicating or distracting from it. One example is found between the aisles of Manhattan fashion retailer Rebecca Minkoff. The store recently installed dressing room mirrors equipped with an AI touch screen, which reads RFID chips embedded in clothing tags, and suggests other items that pair well.

AI’s implementations in retail are varied, and include everything from chatbots that use smartphone camera features to make personalized purchase recommendations, to app-based operating systems that employ customer recognition to streamline checkout. Fundamental to all retail-based AI, however, is its ability to generate a wealth of customer related information for improving, individualizing or even creating entirely new consumer experiences.