Introduction

The Robot Revolution is upon us. A growing number of retailers around the world are using robotic technology to supplement and/or enhance human employees (Grewal et al., 2021). For example, Saks Fifth Avenue is currently employing warehouse robots (by GXO Logistics) to help fulfill online orders (Loten, 2021), while Sam’s Club is using robots (by Brain Corp) to clean its store floors (Repko, 2021). According to a recent survey, approximately half of all retailers plan to employ in-store robots in some capacity by 2023 (Brain Corp., 2021). More broadly, the worldwide retail robot market is expected to grow from $7 billion in 2020 to over $55 billion by 2028 (Coherent Market Insights, 2021). This escalating interest in robot workers has been spurred by a variety of drivers, including advances in robotic technology, human labor shortages, and the need for social distancing in the wake of COVID-19 (Huang & Rust, 2021; Shankar et al., 2021; Subero-Navarro et al., 2022).

The rise of retail robots has also attracted attention from a growing number of marketing scholars (e.g., Grewal et al., 2021; Huang & Rust, 2018, 2021; Noble et al., 2022). Since the Robot Revolution has just begun, most research in this domain is forward-looking in nature and focused on identifying future research directions. For example, Huang and Rust (2018) predict that in the near future, many service-related tasks (such as cleaning and maintenance) will be performed by robots rather than humans. Likewise, Grewal et al. (2021) suggest that research is needed to understand how in-store robots affect shopping behavior as well as how retailers could leverage robots to enhance profitability. In this article, we provide some preliminary insights into these types of issues via an interview with some early movers in this domain (i.e., Daiei and SoftBank Robotics/Iris Robotics). This interview is accompanied by a commentary by Abhijit Guha and Dhruv Grewal (Guha & Grewal, 2022, in this issue), who offer additional insights about retailing’s coming Robot Revolution.

Background

Our interview focuses upon the recent establishment of an in-store robot called the Whiz. This AI-enabled vacuum cleaning robot was initially developed by SoftBank Robotics, one of the largest robotics firms in Japan (www.softbankrobotics.com/jp/product/whiz). To date, over 20,000 of these robots have been shipped to customers across multiple countries, such as the US, Japan, and Singapore (Iris Ohyama, 2022a; SoftBank Robotics, 2020). The Whiz can operate autonomously and be trained to remember multiple cleaning routes in a given store. In addition, its built-in sensors allow the Whiz to avoid both shoppers and obstacles (e.g., products, shelving). The Whiz can operate continuously for several hours before returning to its charging base (Iris Ohyama, 2022a).

In early 2021, SoftBank Robotics formed a partnership with Iris Ohyama (a Japanese home appliance manufacturer) to create a new company called Iris Robotics, which currently handles sales and services of the Whiz in Japan. This version of the Whiz is displayed in Exhibit 1. To date, this robot has been adopted by over 2,500 Japanese firms (Iris Ohyama, 2022b). One of its primary customers is Daiei, a national chain of supermarkets with over 200 stores across Japan (Daiei, 2022). Daiei introduced the Whiz into 79 of its stores in July 2021.Footnote 1 At present, this represents the largest robot deployment of any retailer within Japan. This deployment has attracted considerable attention, as evidenced by Daiei’s recent selection as the recipient of the 2021 AI Retail Award (Retail AI Research Society, 2021).

In order to learn more about the Whiz and how it is being employed by Daiei, we conducted an interview with executives from both SoftBank Robotics/Iris Robotics (Mr. Jun Ikeda) and Daiei (Ms. Maki Totani). Mr. Ikeda is a Director of Business Development for SoftBank Robotics/Iris Robotics and Ms. Totani is the Kanto District Director of Sales Planning for Daiei. We conducted a concurrent interview (via Zoom) with both of these executives on June 9, 2022. Our questions focused on understanding the motives behind Daiei’s introduction of the Whiz, how it currently operates, and how the introduction of this retailing robot has been received by both employees and customers. In addition, we solicited thoughts from its manufacturer, SoftBank Robotics and obtained predictions about the future of robots in retailing. This interview was conducted in Japanese and lasted approximately one hour. With permission of the participants, this interview was recorded and the contents of this recording were transcribed into Japanese and then translated into English by a professional translation firm.

Exhibit 1
figure 1

(Whiz, Iris Edition)

Full size image

Interview

Question

Can you tell us a bit about the robot? How did you begin using the robot? What is the story behind it?

Mr. Ikeda (SoftBank Robotics)

We originally introduced the Whiz (at Daiei) to enhance store hygiene and provide a new customer experience. We first placed it in 18 stores and eventually expanded it to 79 stores. Daiei views this initiative as a way to improve cleanliness and reduce costs. Also, we are using this robot for sales promotion. So far, we have used the robot to promote wine, cakes, and sushi rolls. The levels of both store cleanliness and product sales have improved. A comparison of stores with and without the robot shows that sales are higher at stores using the robots. We have also administered a customer satisfaction survey. The results of the survey indicate that customers and employees are happy with the robot so far.

Ms. Totani (Daiei)

The Whiz was introduced due to increased interest in hygiene management of our facilities following COVID-19. Due to limited staff, our stores needed help with cleaning and sterilization. Also, there is a limit to what can be achieved by humans. This is where the robot became useful. It proved to be an indispensable ally and attracted curiosity. We first demonstrated the Whiz in a store during business hours, which became the initial trigger. In addition to being an effective cleaning system, we believe that the unique appearance of the Whiz has a futuristic vibe. We also found that the robot received a great deal of positive feedback from our customers, particularly children and employees. Since its introduction as a cleaning robot, we have been considering its potential for a broader range of versatile applications. We introduced the Whiz to 79 stores in the Kanto (Tokyo area) and the Kinki (Osaka area) regions.

Question

How long does the robot operate per day?

Ms. Totani (Daiei)

Three hours, on average.

Mr. Ikeda (SoftBank Robotics):

It takes roughly one hour per run (average three runs per day). On average, there are three hours of cleaning, with each store deciding on a time convenient for them to operate, such as 10:00am, 3:00pm, and 7:00pm. Some stores have a shorter period of one hour, whereas others have a longer period of up to five or six hours. It can remember up to 600 routes.

Question

Can you tell us how you are using this robot and what types of outcomes that you have found thus far? Could you describe any challenges in introducing and operating the robot?

Ms. Totani (Daiei)

We have received excellent evaluations in terms of the robot’s cleaning precision.

Mr. Ikeda (SoftBank Robotics)

We have a device that lets us see the relative cleanliness of the floor. On average, using the Whiz leads to a 50% improvement (based on a test that measures the cleanliness of a floor surface). In particular, the robot decreases the number of invisible particles on the store floor.

Ms. Totani (Daiei)

In the beginning, its primary function was cleaning and sterilization. My department was asked if we could use the robot for sales promotions as well. However, we were unsure if it was possible but are currently working on using the Whiz for improving sales promotion. We are extremely focused on this as a department. In addition to improving sales, the robot also boosts cleanliness and prevents crime. We are looking at the satisfaction level of customers visiting the stores and the buzz generated by having robots like this in our stores. Our employees are also pleased to see customers watching the robots with interest. We are working to ensure that this initiative produces a positive outcome. Thus, we are checking what kind of posts about the robot appear on social networking sites, and after seeing some customer postings, our stores started decorating their cleaning robots.

It is difficult to say whether sales promotions are more effective while the robot is in operation. It is challenging to ascertain a causal relationship between the two. It is very complex and there are a lot of things to consider. We experimented several times with sales promotions. One of the first trials was a promotion of Japanese-style confectioneries. We displayed these confectionaries on the sales floor, as well as on the robots via POPs. The POPs included promotional materials and audio recordings. Since the number of units sold vary across stores and time, it was difficult to find a cause-and-effect relation between the robot and product sales. We also experimented with catalog advertising. We loaded the robot with an advertisement that said, “We are now accepting orders for Beaujolais Nouveau wine.” We then compared the number of wine reservations between stores that deployed robots and stores that did not deploy robots. Reservations seemed to increase in the stores that used the robot compared with those that did not. We then asked manufacturers to help subsidize this type of robot-based sales promotion. We first looked at manufacturers that could not engage in in-store promotion (food sampling) due to Covid-19, thinking they may have some left-over budget. We negotiated a subsidy in return for letting them use our robot to promote and advertise their products and conducted field experiments in four stores. These experiments worked very well. It is a trial-and-error process that is still in the early stages with many twists and turns.

Question

What are the incentives for manufacturers to participate in these trials?

Ms. Totani (Daiei)

Well, a manufacturer’s goal is to sell products. The expense is justified if the store can sell more products because of the promotion. Right now, COVID-19 has made it impossible to do tasting samples. However, manufacturers still want to advertise their products, so they are now starting to look for other methods.

Question

How have employees and customers reacted to this robot and how has it affected your firm’s marketing efforts?

Ms. Totani (Daiei)

The store staff felt the robots were like characters. Our company’s mascot is “Mokkun.” So we tried to apply that to the Whiz by giving it a wide face with eyes, a nose, and a mouth. Also, many stores have started decorating their robots in various ways. For example, one store has created an “Ebinya” robot based on a character from Ebina City (a city in Kanagawa-prefecture in Kanto region). They obtained permission from the city to use that character and decorated their local Whiz to make it look like the character “Ebinya” is trundling around cleaning. Several other stores have also begun to take these kinds of steps. Then, there are customers. We see people post about the robot on Twitter and so on. In particular, customers made postings of our robots displaying different promotions.

Mr. Ikeda (SoftBank Robotics)

Social networking sites are already showing actual customer reactions. We receive direct feedback on these sites, with comments such as “This is an amazing age to be alive.” There are many photos and as far as I can see, the posts are almost always positive. We have seen a great number of posts about how the robots are “advanced” and “amazing.” As for whether our employees are satisfied with the Whiz, a recent survey demonstrated that 97% are satisfied and that 95% would like to continue using it in the future.

Question

How have other firms in your industry responded to your introduction of robots?

Mr. Ikeda (SoftBank Robotics)

Daiei was the first company to create an environment in which robots operate in their supermarkets during the daytime on such a large scale. So other companies began to wonder if it was safe to have a robot running around the store during the day. Everyone seems to have a preconception that it would be dangerous and that customers would have to try to avoid collisions with the robots. Surprisingly, these robots avoid people easily, and vice versa. So, competing retailers went to the Daiei store to see what was happening. They were like, “Oh, it’s okay.” You can see that people and robots can really coexist. So they said, “Well, let’s try this at our stores.” Over the past year, the number of companies that have adopted this robot has increased tenfold.

Question

What are your future plans for the Whiz in particular or robots in general? Could you describe new marketing efforts you are planning in relation to the Whiz?

Ms. Totani (Daiei)

I think our company needs robots. This is a major direction for us as a company. We are still in the early stages of using the Whiz to monetize manufacturer-sponsored promotions. So, we are trying to get this part of the business on a favorable trajectory. Currently, we are only doing this in a couple of regions; It should be done on a company-wide basis and we would like to involve as many stores as we can. Our hope is to utilize the robot to generate revenue and improve sales. However, there are some stores where we cannot use the robot, such as stores with narrow aisles or stores with uneven floors or lots of steps. It would be really great if the robots could traverse steps. The existing model struggles to detect even slight differences in surface level. I believe that this is something that needs to be addressed.

We are currently working on in-house standards for promotional materials and are including various other companies in the discussion. We have received some feedback and results from trials, such as “customers are not looking at this material,” or “this material is not being looked at in this location.” During this process, we found that it is better to display promotions at the entrance. Our company has an app, and until now, we have displayed notices in the stores explaining how customers can find information about special offers via the official app. The results revealed that it is definitely better to display this information at the entrance. We would like to use the robot to give visitors information when they enter the store.

I think using the robot to display signage is the most effective method. We had a store where the Whiz rolled around carrying a product. We also received information that other companies are doing something similar. However, in reality, the Whiz is somewhat limited in what it can carry and transport. It can only carry around 10 lightweight products. Right now, we are displaying signage and running a 15-second commercial on a loop. Also, if you go to a different sales floor where another Whiz is running, the same content will be played again. It is easy for customers to understand but may not be engaging for customers who already have an interest in the product. Instead, the robot could automatically play a video when it senses someone nearby. There are also a variety of things that we want to try to promote such as special offers on certain days or offers for different types of members. However, arranging the content has proven extremely difficult.

Mr. Ikeda (SoftBank Robotics)

We could also stop and start the content at a certain time or use “beacon” technology and make use of triggers that run the content when the beacon is touched. It may also be possible to link customers’ smartphones and show different kinds of commercials to those who have installed the Daiei app. There are still some areas that are not technically stable, so we have not yet ventured into these areas yet.

Question

What do you think will be the future role of robots in retail settings? Will they be replacement for humans?

Ms. Totani (Daiei)

This is a difficult question. I do not see robots as being a replacement for humans in any way. Certainly not. Nowadays, robots are becoming increasingly automated and self-reliant. However, when I think about how to make the customer experience quicker and more comfortable, how to make the checkout process smoother, and how to make the shopping experience more pleasant, I think people are superior to robots in this regard. Let me also give an example of a situation in which robots may be preferred to people. When retailers place orders many things need to be considered. Customer demand varies depending on a number of factors such as the weather and the day of the week. Location is another factor. Until now, these activities have been conducted based on human intuition and accumulated experience. The results haven’t been particularly great. The difference in the skill level of each person’s ability has made a big difference in the number of products sold. In order to address this concern, we are currently implementing an AI system for automatic ordering.

Question

Is there anything else that you would like to tell us about these robots?

Ms. Totani (Daiei)

We will only utilize these robots if they are mutually beneficial and if the advantages outweigh the disadvantages. I think this is why we have not deployed the robot for promotion beyond the Kinki and Kanto regions. A few risks still remain. At stores, we have been employing the Whiz for product promotion through playing manufacturer’s commercials. Daiei has been approaching various beverage producers that are keen on this project. There is definitely huge potential and we have been negotiating with manufacturers regarding ways to deploy signage for sales promotions and in conjunction with product placement. We believe many manufacturers will see the benefits of this type of robot-based promotion. Ideally, both our company and our manufacturers are satisfied. Products that don’t sell could start to accumulate, so with new products coming in each week it could be a challenge to make sure everything works. We need to plan for the long term and look a little further into the future. One of the things I think we need to do is to ensure that people will see the robot’s benefits. As a new initiative it is difficult for the robot to be accepted unless it produces benefits and reduces risks. Even within the same company, it could be difficult to get people to understand the robot.

Future research directions

Our hope is that this interview provides a useful portrayal of how robots, such as the Whiz, are currently being deployed in retail settings. As illustrated via this interview, retailing’s Robot Revolution is in its early days. However, the pace of technological improvement in this domain is quite rapid, and even at the current stage of development, retailers can choose from an array of different robots that vary in terms of size, shape and functionality (Huang & Rust, 2021; Repko, 2021). Thus, the Whiz is just one example of a wider robot population. In fact, SoftBank Robotics also markets other robots, such as the Pepper (a humanoid robot designed for conversation and entertainment) (https://www.softbankrobotics.com/corp/robots/). Thus, our inquiry represents only a partial look at a considerably broader (and rapidly evolving) domain and may not be representative of how firms, employees, and/or customers might interact and respond to other types of robots. This subtle but important distinction leads to a number of interesting avenues for future research.

First of all, in contrast to humanoid robots such as the Pepper, the Whiz bears little resemblance to a human (or any other type of living being). In fact, the Whiz looks essentially like a Roomba on steroids. This decidedly non-human-like appearance may be one reason why retailers such as Daiei try to give this robot a more familiar appearance such as decorating it with a face. Prior research suggests that humans are anthropomorphic in nature and may respond more favorably (up to a point) to robots that have a more human-like appearance (Blut et al., 2021; Letheren et al., 2021). For example, Letheren et al. (2021) show that consumers prefer domestic service robots that appear more humanlike in nature. On the other hand, robots that appear too human-like could represent a threat and may be negatively perceived due to speciesism (Schmitt, 2020). Thus, the utility and value of human-like robot such as the Pepper vs. a non-human-like robot such as the Whiz is an intriguing question.

In addition to not looking like a human, the Whiz is also not capable of directly interacting with humans. In fact, from the Whiz’s perspective, humans are simply obstacles that prevent it from performing its duties. Despite its disinterest in (and lack of interaction with) the human species, both employees and customers appear to display considerable fondness towards the Whiz. An examination of social media postings reveal that many shoppers endearingly regard the Whiz as a mascot or a pet. For example, some postings refer to the Whiz as Mokkun (i.e., Daiei’s official mascot).Footnote 2 Thus, even though humans are not directly interacting with the Whiz, they seem quite intrigued by its presence. This observation leads to some interesting questions. For example, how does the ability (or inability) to interact with a robot shape employee and/or customer perceptions of these devices? Likewise, can a robot alter customer behavior (e.g., time and/or money spent in a store) even if it doesn’t interact with a customer? Prior research on the mere presence effect suggests that the simple presence of another individual may enhance arousal and increase drive (Guerin, 1986). Thus, might it be possible that the mere presence of a robot such as the Whiz has a similar effect on in-store shoppers?

Finally, our interview focused on how humans react to robots. This focus makes sense since robots such as the Whiz are largely designed to serve humans. However, as technology in this domain improves, robots are likely to gain enhanced agency and greater autonomy (Repko, 2021; Roggeveen & Sethuraman, 2020). Thus, future research may wish to examine how robots react to humans. This issue will likely take on increased importance as robots expand their role in retail settings. While most humans are likely to treat robots in a respectful manner, some (especially children) may engage in robot abuse (Keijsers et al., 2022; Yamada et al., 2020). To prepare for our interview, we observed several hours of in-store video footage of the Whiz in operation. While this footage didn’t uncover any evidence of robot abuse, it revealed that some adult shoppers appear to intentionally stand in front of the Whiz in order to impede its progress and groups of children sometimes spontaneously gather and surround the robot in a gang-like manner. In effect, some shoppers appear to bully the Whiz. In response, the robot remains still and waits for these shoppers to move on (which they usually do). At present, this type of behavior is unlikely to cause much harm (other than waylaying the robot from its task). However, if robot abuse increases, it is possible that robots may be programmed (or learn) to adopt a more active response, such as asking a bully to move away or reporting him/her to store management. The degree to which these types of robot responses occur and the effect they may have upon store operations, customer experience, and robot welfare are intriguing questions for the future.

Exhibit 2
figure 2

(Whiz deployed at Daiei)

Full size image

Conclusion

The rapid development of new technologies such as AI and robotics has captured the attention of both marketing scholars and practitioners. As noted by Fournier and Mick (1999), new technologies typically entail a number of paradoxes and often force humans to adopt new practices, which may be both good and bad. Thus, it is not surprising that the coming Robot Revolution is attracting anticipation as well as trepidation. While it is difficult to predict the future, our early look at Daiei’s deployment of the Whiz suggests that the future of robots in retail may be more satisfying than terrifying. For example, our interview reveals that the presence of the Whiz engenders positive affect from both employees and customers and that it supplements, rather than replaces, humans. In addition, initial indications suggest that the use of the Whiz as a promotional device (as shown in Exhibit 2) is generating positive returns for Daiei and presents an unexpected use case for SoftBank Robotics/Iris Robotics. Indeed, the improvisation of the Whiz from its intended use as a cleaning device to its new use as a promotional tool is surprising and suggests that robots may be used by retailers in ways that both practitioners and scholars may not predict. This uncertainty will likely make the development of appropriate theories and concepts particularly challenging. Thus, we encourage future scholars to follow our lead by observing and conversing with robot manufacturers, users, employees, customers and perhaps even the robots themselves. In particular, inductive-based approaches such as grounded theory (Goulding, 2005) and theories-in-use (Zeithaml et al., 2020) may be particularly valuable. We also encourage scholars and practitioners interested in this domain to read the prescient commentary to our interview offered by Abhijit Guha and Dhruv Grewal (Guha & Grewal, 2022, in this issue), which directly follows our article. In closing, we hope that this look at rolling out the Whiz provides some early and valuable insights into retail’s coming Robot Revolution.