Saying is One Thing, but Doing is Another. The Case of a Little Robot in the Bank.

By Kristina Cigic, CMRP

At ATB Financial, we’re testing out a new team member. She’s about 4 feet tall, her name is Pepper… and she’s a robot.

Though people will play a key role in ATB’s relationship with its customers, there is so much opportunity with robotics and artificial intelligence to learn how we can create capacity for our team members to focus on the best part of their work—customer relationships.  Let’s face it, with the pervasiveness of mobile and online experiences, the way people want to experience banking is changing. The question is whether a little robot like Pepper (and the big software behind her) can add value to clients and team members.

The preliminary research phase over the past few months has involved Pepper touring three of Calgary’s branches, each chosen for different reasons. All had high foot traffic, but some were more ‘traditional’ style branches whereas others were more ‘open concept’. Each branch has a different core profile of customers, whether they differ by age, profession or lifestyle.

While touring, Pepper’s main functions were to greet and entertain by dancing with customers or taking a quick selfie, and recommend a range of financial products. People can either speak with Pepper directly, or choose to interact with the tablet on her chest instead.

The research questions were simple enough: What effect will Pepper have on how people think about ATB? Can she attract people to the branches? How will people respond to a robot in the branch? What will the interactions look like and could people trust Pepper with their banking matters? Ultimately, what role can Pepper really play?

Good research is the heart of creating any great customer experience. UX research is not only about gauging people’s perceptions and attitudes about something; it’s about understanding and making inferences about their behaviour. It’s about realizing that what people say and what people do often contradict each other, and that we need to dig deep to uncover the truth. In order to answer our research questions, our methodology had to incorporate ways to evaluate the attitudes and behaviours of our customers. We couldn’t just ask people what they thought; we also had to observe them in a real-world context to truly understand the role Pepper could have in their lives.

We spent around 30 hours doing ethnographic observation in each branch, watching team members and customers. We sat back and recorded anything we saw that seemed relevant: the time of day, group size, whether it was paycheque-getting time or bill-paying time in the branch.

Don’t get me wrong, we did a lot of asking as well!  We talked to team members to get their observations and see what comments customers were making. We intercepted people who chose to have a banking-related interaction with Pepper (and not just an entertaining one— though there were plenty of those). And we did some traditional quantitative research, as well. We sent out a survey to all of our customers who visited the branches while Pepper was there and commissioned a short study to gauge awareness of Pepper and brand impact among the general population of Albertans.

What we found validated the methodological approach we had chosen. As it turns out, what people said was not the same as what people did!

Through our quantitative research, we found that customers did have some trouble trusting Pepper in meaningful banking roles; they preferred to see her do things like entertain or inform people about branch events, rather than use customer data to make recommendations. But overall, the majority of people we asked said they think robotics has a place in the future of banking and that Pepper adds a ‘cool’ factor to ATB. A vast majority of those who interacted with Pepper in the branch claimed they enjoyed their experience and that the interaction was easy for them.

Now the observational data plays in – we found that, although people seemed to notice Pepper, many chose not to interact with her and those who did seemed uncertain around her. It was common that people would abandon their interactions early.  Those who engaged with Pepper tended to be more content with lighter interactions, like seeing Pepper dance or take a selfie. Not many were interested in having Pepper talk to them about banking related needs.

If customers told us that Pepper was ‘cool’ and they enjoyed interacting with her, then why weren’t those interactions more frequent? Why weren’t they more meaningful? When we looked at all of our data sources, a fuller picture began to emerge. These are some of our conclusions:

  • In Pepper’s case, generational differences were important. In her current state of programming, Pepper is limited and her demeanor is not always smooth. Millennials tended to have high expectations, but because Pepper isn’t conversational (or human) enough yet, those expectations were rarely met. On the other hand, Boomers seemed to have lower expectations. They would take more time to speak slowly or repeat themselves if Pepper needed them to. All of this impacted the depth of interactions people had.
  • We also concluded that a branch location may just not be the right place for Pepper. When customers take time to physically come into a branch, they’re usually doing so for a reason and they want to finish their business quickly. It may not be the right environment for a meaningful interaction. For instance, we noticed that when there was no line up for service, people would walk right past Pepper and up to the teller. Many people also mentioned that they just didn’t want to miss their turn in line by being distracted with her.
  • Social factors were important. Robotics is new in the everyday world, and people don’t want to look silly talking to Pepper! One of the things that led us to conclude this was that people were much more likely to interact with Pepper when they were in groups or if someone was already interacting with Pepper. It was much rarer that someone alone in the branch would approach our robot friend.

There is so much potential for ATB’s new robotic team member. We know that because our customers have told us that; however, our method reinforced the fact that good UX research really does need to balance the ‘saying’ and the ‘doing’.

This first phase of research uncovered some of the fundamental barriers to having Pepper in our branch network but we are still researching Pepper and assessing her role in public places like universities, show homes and registries. What more can we learn about what she brings to our in-branch, corporate office and in-market experiences? To keep up with the work we’re doing to transform banking visit We’re excited to share our journey.

Kristina Cigic, CMRP is a Senior UX Research Manager at ATB Financial in Calgary, and also leads as a member of the MRIA’s Alberta Chapter Board and the ELTF. Credit to Tiffany Kieboom, Sr. UX Research Manager at ATB, who crafted the methodological approach.

Leave a Reply