Creating a virtual assistant for effortless financial services
When it comes to using virtual assistants (VAs) in banking customer care, financial brands are pushing the technology envelope. Many notable institutions have invested in, and are rolling out, VAs (also commonly referred to as conversational bots or chatbots) with the goal of providing effortless customer service to millions of consumers at a time.
Customer expectations for prompt service, a growing comfort with automated support and brands’ dedication to improving the overall financial services customer experience are some of the biggest contributing factors for an increase in virtual assistants. But while the solution sounds promising, where do financial services companies new to the world of AI and bots begin in their digital journey?
The rise of virtual assistants
A good philosophy to have when approaching the creation of a virtual assistant is “pain points first, products second,” says Deniz Güven, global head of client experience at Standard Chartered Bank (SC). Güven is behind Standard Chartered’s Stacy, a VA created to streamline the customer experience for SC’s large, global, multilingual customer base.
Güven emphasizes that en effective conversational bot can only be developed when a brand understands — in detail — the key service challenges their customers face on a regular basis.
Tae Moon, senior manager of emerging capabilities at TD Bank agrees. Moon is one of the key executives at TD Bank involved with the planning, creation and implementation of Clari, a conversational AI-based virtual assistant that provides service to TD customers.
He says that because AI technologies are increasingly versatile and far-reaching, defining clear use cases and applications from the beginning is important so financial brands can stay focused on building the right functionality into their chatbots.
Standard Chartered Bank sought to fully grasp and contextualize its customers’ usage habits before creating its VA. “What we found was 87 distinct pain points that helped inform the design, value proposition and branding of our chatbot,” explains Güven.
The user interface (UI) of a bank’s mobile app, for example, might force customers to click through, swipe and enter information to complete a simple money transfer. Güven says that UI complexities like this helped SC design a virtual assistant that would simplify usage.
Six steps for conversational bot success: A helpful primer for chatbot implementation
Make the most of your conversational bot investment with our easy-to-follow guide featuring best practices that can be applied to your digital transformation journey.
Creating a virtual assistant
Once your customers’ pain points have been identified, the next step is collecting and applying the right data in order to design and create a natural customer-brand interaction.
“Once there is a clear use case and application, focus should be directed to the availability of — and accessibility to — customer data that will ultimately power the VA,” says Moon, who adds that brands should take stock of the large amounts of customer data they already have to give their VAs a “head start”.
Güven, meanwhile, recommends that brands view VA design in terms of three key aspects: Enquiry Services, Transaction Services and Conversational Services. These three areas cover a wide range of functionality, from simply providing an account balance when asked, to having a natural-feeling conversation about complex investment topics.
“When creating a VA it’s important that the conversation feels as natural as possible,” he says. “While creating Stacy, for example, we created a huge library of natural-sounding phrases, words and sentences she could draw from in various interactions.”
Roll-out and continuous optimization
For Moon, the successful introduction and implementation of any AI-powered virtual assistant centers around managing customer expectations. VAs can be extremely powerful for certain tasks, but tend to fall short in others. Banking customer care interactions that entail complex documentation and/or identity authentications (a mortgage loan approval, for instance) are still best handled primarily by a human support agent.
“To properly manage expectations, TD has employed various tactics including a phased rollout, clear messaging of Clari’s capabilities and active communication with pilot users to manage the collection of feedback and adapt quickly to resolve any issues in the pilot test phase,” he explains.
For Güven at Standard Chartered, one of the continuous improvement and optimization goals for Stacy is to increase the portability of the VA. The bank is exploring ways to extend the functionality of Stacy beyond just a banking app or website, to other platforms like Facebook Messenger and WhatsApp.
“Portability means taking a VA like Stacy and giving it the means to follow the user through other aspects of their life. Even if I don’t open the Standard Chartered app, we’re working to make Stacy accessible to provide conversational service related to that customer’s financial life via other channels,” Güven says.
In the future, Moon advises that financial service brands take a close look at how their virtual assistants can (and should) become further embedded in Internet of Things (IoT)-enabled devices.
“Whether our customers order lunch through [Amazon’s] Alexa, buy gasoline through their onboard car system or authorize an automatic purchase of milk through their smart refrigerator, all these interactions require financial transactions and insights,” he says.
Giving customers more choices
As technology evolves, Moon’s goal with Clari is to give customers the same experience regardless of what service channel they choose, whether it be in person at a local branch, over the phone, or through digital channels.
In the end, Güven urges banks not to overlook the obvious when designing a VA for their customers. Chatbots should address specific customer pain points, make use of existing customer data and most importantly, provide friendly and effortless service.