As we handle more of our daily tasks online or by mobile, banking queries have struggled to match the pace of information exchange in other areas of life. Conversational AI makes it possible to offer a level of responsiveness and personalised services that aren’t financially viable to offer at scale with traditional advisors and contact centre staff.
One of the biggest challenges banks and their customers face is responsiveness, with customers often waiting in a queue for hours seeking for help. It’s such a problem that almost two-thirds (63%) of customers say they would consider switching banking providers if communications don't meet their expectations.
Regulators are also taking note of what they’ve assessed as ‘poor banking customer services’, with The Central Bank of Ireland stating it will apply pressure on banks to improve. In the UK, the FCA have noted how digitalisation improvements in origination are failing to extend to customer services, launching its new Consumer Duty act. This states that financial services providers must be able to make processes and services that are as easy to change or cancel services as it was to buy the products in the first place.
The good news is that Conversational AI technology has now matured to safely enable banks’ traditional trusted advisor relationship in new settings: whether it’s on the phone, on messaging platforms, in a secure webchat, email, or across any combination of channels. The much vaunted ‘omnichannel experience’ that has been the goal of marketers can be improved upon to serve customers, providing proactive support and guidance to customers as well as handholding when needed.
Without the valuable personal touch of an advisor, individual customers all get corralled into commodity-based processes that may or may not be directly relevant to their needs, and they feel it. Conversational AI allows you to provide tailored services.
While it would be impossible to scale up staffing to the size needed to provide a personal financial advisor whenever and wherever your customers would like help, Virtual Agents (VAs) are easily scaled and allow customers to interact with your institution in the same ways they would with any normal member of your staff. Well-implemented VAs enable customers to access self-services instantly via any channel, independent of backend system silos or normal business processes, resolving requests immediately.
This is more than a change in the technology mix, it will mean a fundamental shift in the ways banks interact with customers: from reactive to proactive, from restricting service based on perceived value to extending services for all.
Notifications that enable action
For example, if a customer has missed a payment, you could send them a message letting them know they’ve fallen into arrears. If that message also leads them to an online payments link, they’ll be able to act on it when it suits them.
While we can automate these simple reminders, Conversational AI allows us to handle more complex cases where additional information affects outcome. The fact is, while systems can ‘know’ that something has occurred (such as a delinquency), those systems still don’t know the reason until we initiate a discussion. Perhaps the customer has suffered a setback that qualifies for deferment or some other allowance. To get the customer back on track we need to gather more information from them so that we can provide guidance and support.
This is where the benefits of Natural Language Understanding (NLU) really shine. Using this technology, we can create systems that ‘understand’ different types of reasons, record them in your backend systems, recommend actions such as seeking a deferment, and then help customers throughout the entire process of uploading documents and filling out forms. This support can be provided as a text or voice-based encounter or a combination of both.
People often prefer to provide information about a setback without telling a ‘real’ person about it. VAs can help mitigate this issue; if a customer tells us about a hardship, the VA response will be sensitive to their situation, creating a more comfortable user experience.
Guidance along the way
Everything described so far is still somewhat reactive because it has been triggered by an event. What about truly providing proactive assistance? Conversational AI can provide advice on what to prioritise based on product information (interest rates, late fees and so on) as well as relative importance (mortgage payment probably trumps credit card bill for instance).
Beyond calling your customers by name, Conversational AI can carry over memory between sessions so it can recognise why a customer might be calling instead of forcing them to re-explain themselves each time. Going back to the example of the document, when the customer calls, the VA would greet them by name and might say: “When we last were in contact two days ago you were in the process of applying for a loan deferment. There is a document due before your request can be processed. Would you like to upload it now?” This provides an actionable prompt and may save the customer having to explain why they’re calling.
Realtime or anytime
Calls demand focused attention and Conversational AI can be built to deal with unplanned interruptions. A caller can ask to switch to text-based channel to continue the same interaction as and when they find time to respond. They don’t need to worry about calling back, but if they do, the VA can pick up where they left off.
Even when a VA reaches a point where it’s not yet able to fully support the customer, it can continue to help in the background.
If the interaction gets handed over to an employee, the conversation history is waiting for them, so customers don’t need to start over.
Even with banks’ expanding self-service capabilities, customer wait times is key. Virtual agents scale, so that most requests can be handled without delay, even during peak times.
The burden of keeping on top of financial obligations and opportunities needn’t fall solely on the shoulders of already time-stretched customers.