Sales robots: when less means more on the sales floor
Published on 03/03/2026
When a sales robot heads over to a customer in a store, the interaction can easily backfire. Shoppers often regard the approach as intrusive and end up walking away. A series of experiments carried out by NEOMA researcher Xingming Yang and her colleagues show that robots must rigorously respect a consumer’s personal space. They also need to keep a low profile, respond only when solicited and leave the decision to engage (or not) in the hands of the shopper.
Robots promise a smoother shopping experience on the sales floor, guiding customers, answering questions and recommending what to buy. Yet this simple‑sounding innovation may have unintended consequences. While mobile robots are capable of acting like salespeople and can approach shoppers to initiate a conversation, they risk triggering an avoidance reaction in customers.
This paradox — that the most “helpful” technology is not always the most effective — lies at the heart of the study by Xingming Yang and her co-authors. The researchers ran a series of controlled experiments where they systematically adjusted several key variables, including the robot’s physical proximity to customers, its level of initiative and its degree of autonomy. In addition, they measured how much intrusion customers felt they had and whether the presence (or absence) of a human salesperson changed the outcome. The results generated a list of guidelines for designing and deploying in-store robots: they should be less “pushy”, maintain a lower profile, respond only when approached and — the key finding — let customers decide whether they want to engage or not.
Give customers space
In retail, approaching customers is part of the standard commercial “script”. For shoppers, however, it can be a mixed experience: if the salesperson comes too close, it can feel so intrusive that some consumers report they dislike being shadowed while browsing.
In these circumstances, a mobile robot that heads towards a customer creates an extra layer of uncertainty. Its approach is often harder to anticipate and is more likely to intrude on the shopper’s personal space. The authors draw on proxemics, the study of spatial relations between individuals, together with the idea of personal‑space invasion, to explain why this simple movement may trigger a sense of intrusion.
All the study’s findings point to the same conclusion: a robot that approaches a customer is perceived as more intrusive than a human salesperson doing the same thing. This encroachment spoils the shopping experience and may prompt avoidance behaviours, such as choosing not to ask for advice or feeling the urge to leave the store.
A good robot is a low-profile, responsive robot that is easy to switch off!
The study found that there are straightforward alternatives to the model of the “proactive” robot that approaches customers. For example, shoppers respond more favourably to a stationary in-store robot that does not initiate contact than to a robot that imposes itself, which in turn enhances the customer experience. There is another option: a reactive (rather than proactive) robot that does not approach customers and interacts only when asked. This model boasts the same advantages as the mobile robot without eliciting feelings of intrusion.
Furthermore, the researchers show that giving customers the choice to interact leads to better outcomes. When participants are given a clear opportunity to opt out, they not only feel less invaded but also report higher levels of satisfaction, meaning they are less likely to feel the urge to head for the exit.
It follows that, to create better experience, sales robots should not seek to imitate salespeople or impose themselves but offer consumers a measure of control.
How to deploy robots effectively: a user's guide
In the authors’ view, the implications for management are straightforward: if a brand encourages interaction by programming a robot to act like a salesperson, it should expect counter-productive reactions due to a sense of intrusion. Accordingly, the researchers offer a recommendation: rather than mimicking human scripts too closely, brands should develop their own voluntary modes of interaction that leave the choice in the customer’s hands.
The study additionally shows that having a “human safety net” can make the experience feel safer, even though the presence of a salesperson does not diminish the perceived intrusiveness when a robot approaches shoppers. By contrast, the best setup in terms of customer satisfaction is when the robot remains stationary and a salesperson is available if needed. While humans help ensure that the interaction runs smoothly, the decisive factor is the robot’s behaviour, which requires careful fine-tuning.
Finally, the authors argue that the technology should be rolled out in stages. They specifically recommend providing clear information about the robot’s capabilities and maintaining a sense of synergy with employees. Brands should also start by phasing in reactive robots that customers can choose to interact with so they get used to the technology. In passing, the research team also notes a potential reputational risk: replacing humans with robots may have an adverse impact on a brand’s ethical image… just one more reason for prioritising low-profile, “on demand” robots.
Find out more
Yang, X., Garnier, M., & Djelassi, S. (2025). “Robot, don’t approach me!” When a robot emulates a salesperson approaching in-store: Impact on perceived intrusiveness and behavioural expectations. Recherche et Applications En Marketing (English Edition), 20515707251379970. https://doi.org/10.1177/20515707251379970
Related news
Professor

YANG Xingming
Xingming Yang is an Assistant Professor of Marketing Management at NEOMA Business School. Xingming specializes in behavioral studies and works with other research methods such as qualitative interviews. Xingming's research focuses on consumer behavior in the digital environment, as well as human-rob