Implementing Chatbots in Financial Services Implementing Chatbots in Financial Services
Many firms begin their foray into chatbot technology by replacing or augmenting internal service desks. Functions such as information security password resets, IT software installation requests, and HR inquiries are typically the first to use chatbots as the dialogs are simpler and more limited in scope, with a small number of backend systems with which to transact. These firms then turn to deployment in more complex, external-facing capacities, such as customer service, to reduce the number of calls requiring a human agent, thus reducing cost. However, if chatbots are not properly implemented or trained, customers can get easily frustrated as their requests are repeatedly misinterpreted.
In order to maximize the benefit of a chatbot implementation, the range of typical customer dialogs must be compiled and documented, a complete dictionary of industry and company specific terms assembled, and the corresponding back-end systems and transactions identified.
Sentiment engines should be leveraged, such that if a customer is getting agitated, the call is immediately transferred to a human representative who can process the request. In a similar manner, if the customer is deemed to be highly satisfied, the call might be transferred to a representative where additional products or services can be offered.
Direct customer-facing chatbot deployments are not the only option. The technology can be used by human customer service representatives to provide expert guidance or as an accelerator in accessing required transactional systems. For example, chatbots can “listen in” to a customer service conversation from the representative’s side, never visible to the customer. Chatbots can guide the representative in addressing the customer’s issue, accelerating the session by accessing the company’s various transactional systems to provide the required information, or impart the updates necessary. With the aid of the “expert” chatbot guide, the duration of the call can be shortened, allowing more calls per representative, thus reducing expense.
With less wait time and a superior service experience, customers are more likely to be receptive to suggestions for additional products and services. Here again, chatbots can function as a guide, using ML to suggest offerings that are likely to appeal to the customer. With full access to a customer’ transaction history, demographics, account profile, and product set, the AI-enabled chatbot can make recommendations in real-time, without delaying the call.