Conversational Banking Vision

Challenge

[Large Nation Bank] was a latecomer to natural language processing (NLP) and virtual assistant (VA) technologies but aimed to leapfrog its peers. My role involved developing a vision for how the bank might interact in the future with customers.

Approach

Research

I took over this project after initial SME interviews, which, unfortunately, provided little data on customer behavior, so I engaged the bank’s CX team. Their research showed a substantial shift in customer channel preferences from calls to digital channels — specifically SMS, not email.

While researching the bank’s SMS initiatives, I was made aware of a Rich Communication Service (RCS) proof of concept the innovation lab was working on. I received a demo where I heard that Soul Machines was running a virtual human POC for the lab.

With this information, I studied RCS analogs and how mobile apps like WeChat were moving into the financial transactions space. On the subject of virtual humans, I investigated the uncanny valley phenomenon and how endearment balanced anthropomorphism.

 

Pepper’s big-eyed adorableness keeps it out of the uncanny valley.

Synthesis

From the outset, the business focused solely on the competition and pushed a text-based VA solution. I created a market landscape that included Bank of America’s Erica and Capital One’s Eno to satisfy their interests, but I also directed them to a better opportunity.

Whether a simple, text-based autoresponder or a physical robot, VAs were no more than chatbots with an interface predicated on the environment. The opportunity was to give meaning to the interaction and extend the customer relationship.

I researched how individuals form personal relationships with technology, such as virtual girlfriends like Xiaoice. I looked into how the bank used data to generate behavioral probability models to drive predictive banking.

In my POV, I laid out the factors that cause people to suspend disbelief, accept a cold interface as an empathetic friend, and even develop an emotional attachment. I recounted the effects of notification addiction and how burnout drove people to reject technology. I also referenced the discomfort people feel interacting with systems that eerily know them.

 

Research pre-reads immersed stakeholders into the opportunity space without having to parse lots of information.

Exploration

Next, I led a workshop of 30+ stakeholders and strategists, colocated on both coasts, and assisted by four table leads. The challenge: Invent ways to extend the bank’s relationship using non-humans in Year 2030.

The participants conjured ideas about the guises AI might take and then demonstrated how these might work by creating journey maps based on the plausible scenarios I provided. Post-workshop, I worked with the project stakeholders to prioritize the solutions based on feasibility and business interests.

Using the prioritized concepts, I developed a visual narrative that told the story of an AI-based avatar who helped a family upscale their home to accommodate their immigrating grandmother.

 

Participants wanted to give AI both a face and a personality.

Evaluation

I composed a protocol based on issues I had uncovered during research, in addition to validation and value prop questions. A market research firm was engaged to explore the narrative with three focus groups comprised of emerging-affluent, mixed-generational participants.

The study revealed contradictory motivating factors. Customers liked the cute animated avatar, especially for simple day-to-day banking matters. However, when it came to major decisions such as loans, participants wanted the assurances afforded by a human.

Proactive banking had mixed reactions, too. Some found an intervening entity would save them from themselves, while others associated it with the creepiness of highly-targeted ads. Trust was a motivating factor.

 

Meet Paulo, the animated virtual banker and one of many personalities our fictional grandmother, Mary, can choose.

Outcome

Our commercial-like videos are the deliverable most prized by our sponsors. So, after testing, I compiled the most compelling parts of the narrative and met with our video production partner. The team produced a script, scouted talent, and secured locations. But before they could start shooting, the COVID-19 pandemic forced them to shut down production. After a year, the sponsor canceled the video but moved forward with the product strategy I’d developed, which I’m not at liberty to reveal.