Using Virtual Assistants to Drive Consistent, Multi-channel Customer Experiences

By Brad Stewart, Head of NLP, AI Enterprise Solutions, Wells Fargo[NYSE:WFC]

Brad Stewart, Head of NLP, AI Enterprise Solutions, Wells Fargo[NYSE:WFC]

Virtual assistants have become commonplace in modern society, with the wide prevalence of Siri, Alexa, Google Home, and others. Many of us have come to rely on these assistants to tell us the daily weather, turn on the lights as we enter a room, and play our favorite podcasts. We’ve been using voice for thousands of years to interact with each other, so it’s a natural evolution that we interact with our technology this way versus always tapping and clicking a screen.

In addition to providing a more natural way to communicate, virtual assistants make it easier for customers to engage with a brand 24/7 from wherever they are. Whether it’s from a mobile device or computer, through various apps and messaging platforms, or even through your remote control, virtual assistants make it easy for customers to find what they’re looking for. They can often answer a wide range of questions, help consumers accomplish tasks, and provide relevant and personalized guidance. As the convergence of data analytics, APIs, and voice-controlled devices continues, there’s opportunity for banking and other services to become seamlessly integrated with a consumer’s daily life.

Because of the huge amount of potential to help drive a transformative digital experience for our customers, Wells Fargo is working to build out natural language processing (NLP) capabilities on our own platform to give maximum flexibility and control over future capabilities and experiences. However, there are many third parties in the market offering products for companies who want to leverage virtual assistants but are not able to invest in the talent and resources to build and maintain their own platform.

Humans Still Matter

Though virtual assistants have been growing in sophistication and learning new skills, there are many situations where a customer may want or need to interact with a human, and the best customer experience requires a seamless hand-off from bot to human.

Over time, we plan to introduce virtual assistant capabilities across a variety of channels where our customers may want to carry out banking tasks. We are taking a “single bot” approach, allowing for a consistent experience for our customers no matter where they communicate with us. The bot will be able to answer questions and perform tasks in the same manner, regardless of how or where the customer interacts with it. And when conversations need to be transferred from bot to bankers in any given situation, the banker will have access to past conversations to essentially eliminate the need for the customer to explain their question or issue multiple times.

Additionally, the next wave of NLP development includes enabling virtual assistants to better understand tone and emotion. For example, if a customer is using words that may signify anger or frustration, we can get them connected to a real person faster, as customers are looking for an emotional connection to the brands they interact with—and humans still do that best.

In addition to providing a more natural way to communicate, virtual assistants make it easier for customers to engage with a brand 24/7 from wherever they are

Team Member Support

Virtual assistants aren’t only good for customer interactions; they also have applications to help our team members. For example, we have extensive content on internal websites across a wide range of topics such as parental leave policy, education reimbursement, international travel requirements, tips for new managers, and much, much more. Navigating through so many sites with all this information can be challenging, and traditional keyword search is not always effective. But using AI and NLP techniques we can train a virtual assistant to understand the concepts contained in all this content, allowing team members to ask natural language questions and find answers quickly and easily.

Security and Privacy

While voice is natural, there are obvious situations where you wouldn’t want Alexa or Siri reading out information, and finance is full of these examples. Voice can certainly be a valuable biometric in confirming one’s identity, but helping customers keep information private as voice technology becomes more dominant is an important ongoing discussion. It’s likely we’ll see a mix of voice and touch screen prompts that work together as an early solution.