Pepper - Shopping Assistant

Pepper is a humanoid ( robot that looks like humans) robot manufactured by SoftBank Robotics. Pepper is powered by AI and is designed to resemble and mimic human features and behaviours to enhance its ability to interact with people. It was introduced in Tokyo on June 5, 2014, in June 2021 SoftBank paused production of Pepper, citing weak demand. The cost of Pepper is approximately ( 44,000 USD) and requires extensive maintenance.

My Role: Lead Researcher

My Tasks: Questionnaire generation, conducting qualitative research with the team, data analysis and synthesis, final report creation and final presentation to the management.

Time Frame: 3 Weeks

How we prototyped Pepper?

For the research project, we built a prototype of pepper using thermocol blocks that are typically used in construction and attached a tablet to the chest of the prototype for software interaction. Finally, for the movement, we fitted a battery-operated RC toy car under the legs and controlled it using a remote control. After a few iterations and floor tests, the prototype was ready for testing.

The actual pepper robot
Image credit: https://en.wikipedia.org/wiki/Pepper_(robot

The prototype we created using thermocol

My team members who build the brilliant pepper model using thermocol.

Business Case

One of the known retail business brand in India was keen to deploy Pepper, the humanoid robot in their stores to enhance customer experience and build a physical shopping assistant who can tell customers about offers, help find product location/ aisle, tell them about the stock availability , product ingredient details etc . Though the idea was super exiting, the cost involved to procure pepper and to build a software layer that integrated with SAP ERP backend and UI layer was challenging and expensive to achieve. Along with my team I decided to run a social validation experiment with a prototype to check if the concept was worth investing.

Qualitative Research Questions

The first part was to plan what we wanted to achieve out of the research exercise and the data we wanted to collect that would help gain insights into the customer experience if Pepper were to serve as a shopping floor assistant.

I as the research lead created the first draft of the questionnaires and collaborated with my team members who were working as design interns to refine it further. Three of us, including myself, formed the research team who would interview customers and collect data on the shop floor.

As this was more of exploratory research, to collect as much data as possible and I engaged a qualitative research method / questionnaires . Below are the 10 qualitative questionnaire that we finalised.

What are customers' initial reactions and emotions when encountering Pepper in the retail store?
How comfortable are shoppers with approaching and interacting with Pepper?
How do customers perceive the usefulness and effectiveness of Pepper in assisting with product location, availability, and details?
To what extent do customers trust the information provided by Pepper regarding product details, including price?
How does the presence of Pepper affect customers' overall shopping experience in terms of engagement, convenience, and satisfaction?
What are customers' expectations and preferences from a humanoid robot like Pepper in a retail setting?
Do customers perceive Pepper as a practical tool for enhancing their shopping experience?
Are there any concerns or reservations customers have regarding using Pepper in a retail store envoirnment?
How does customers' age, gender, and technological literacy influence their perceptions and acceptance of Pepper in the retail environment?
Would customers be more inclined to visit a retail store that employs Pepper over stores that do not, and why?

Collecting Research Data

We placed the pepper prototype at the entrance to gain insights into shoppers’ first expressions and curiosity. For those who were curious and wanted to touch and interact with pepper, we asked them if they were comfortable answering a few questions regarding pepper.

This is how we placed the prototype at the retail store entrance to understand the reactions of shoppers who entered the retail store. The observation method was used to take notes and capture some of the shopper's emotions and behaviours.

The prototype model was also placed in different shopping aisles. We used the remote control to move the prototype near shoppers who were either searching for products or were looking confused. We wanted to check if the customers were comfortable with Pepper going to them while shopping and proactively proving help and being a shopping assistant.

A product scan function was added to the tablet on the check to show the value that Pepper can add during shopping. Though it was a hardcoded prototype, when a shopper scanned a specific product, it provided nutritional details, price, and offers, including the product feedback option. Here we wanted to test whether consumers were comfortable interacting with the tablet on Pepper including shopper's emotions and feelings. This was recorded using a 1:1 questionnaire method.

Lastly, we also placed Pepper near the POS ( Point of Sales) to capture shoppers' experience and if made sense to have Pepper as a shopping assistant.

Conclusion

Based on the observations and customer interviews, the conclusion is that deploying Pepper in the specific retail store environment requires a huge investment. We couldn't calculate the ROI of such investments, and the shop floor layout was not ready yet to handle the humanoid movements. Here are some of the key reasons supporting this conclusion:

Cost and Complexity: The cost involved in procuring Pepper and building a software layer to integrate with the SAP backend was deemed challenging and expensive.
Social Acceptance Concerns: While initial reactions were positive, and there was a notable curiosity on the shop floor regarding the presence of Pepper, both adults and kids, it's important to mention that kids seemed to have a particular affinity towards the humanoid robot and wanted to play with it while the parents shopped indicating a less appeal from the adults.
Practicality and Usefulness: While customers perceived Pepper as a good-to-have option, the research data indicted that shoppers preferred to ask a shop floor staff for assistance or a self-service kiosk during shopping rather than asking a Robot.
Technological Readiness: Pepper's software platform was challenging to integrate SAP solutions, indicating that the technology may not be sufficiently mature at the point when this research was done for effective deployment in a retail setting.

Considering the challenges and limitations identified, it is prudent to conclude that deploying Pepper in the retail store at this stage may not be the most feasible or practical solution