The greatest fear for banks in the changing landscape of financial services is to lose grasp of the customer relationship and become utility commodity providers, or “dumb pipes”. This fear is strengthen by changes in consumer behavior and expectations, regulatory changes such as PSD2, new business models and new technology. Artificial intelligence is one of the technological paradigms that is set to transform financial services in the years to come.
How will banks stay relevant in a perfect storm of changes affecting nearly every aspect of the industry? According to Accenture, banks must become an indispensable Everyday Bank to meet changes in consumer behavior and expectations.
Everyday banking is more than providing a mobile banking app showing a list of transactions and the ability to transfer money. In order to be relevant in the customers daily life banks need to offer something beyond banking. Common characteristics of everyday banking solutions are a mission to improve financial literacy and challenge users to become aware of spending behavior. The first generation of everyday banking advice utilize a combination of data analytics based on spending behavior, gamification and behavioral psychology to achieve this.
When it comes to digital banking, the smartphone is a key driving force, and according to GSMA, AI is the next battleground in smartphone innovation. Google is aiming to make voice-based artificial intelligence a much bigger part of how people interact with their smartphones in Google Pixel.
When we are moving closer to the post-app era, the transition from conventional graphical user interface to a conversational user interface, a term for online businesses that’s powered by natural language technologies we need to rethink the approach to everyday banking services. With a combination of rich visual interfaces and artificial intelligence (AI) technologies from Google, Facebook, Amazon, and others, brands can scale relevant, personal, and helpful interactions with customers.
The first step on the way to virtual financial assitants is the introduction of chatbots. Banks are already experimenting with this.
SwedBank has applied a chatbot on their website that answers all visitor questions, and challenger bank Atom offers a virtual agent who gets smarter with each conversation, and recently, ABN Amro showcased a prototype of a chatbot that tells you what your mortgage limits are after you answer a couple of questions.
By transitioning from graphical UI til conversational AI, a majority of banking services could be automated through simple voice or chat request like “what is my daily spending limit until my next paycheck” or “approve and pay my outstanding bills”. The impact of virtual assistants is not limited to simple daily banking requests. AJ Bell is aiming to utilize Facebook chatbots to offer trading services for investors, as well as Personetics that delivers personalized financial guidance to customers everywhere through their AI-powered chatbot.
Virtual financial assistants present an unprecedented opportunity for incumbents to increase customer engagement. Banks are in possession of vast amount of user data that is invaluable for unsupervised machine learning as well as intangible know-how through years of experience for supervised machine learning.
Instead of struggling to find the best financial products for our needs, we will now be tasked with identifying the right AI tools to manage the task for us. Or it is plausible to imagine that AI-powered fiancial advisors wil become virtual subcontractors to your virtual personal assistant for everything else like Facebook’s M, Microsoft’s Cortanta, Siri from Apple or Amazon’s Alexa. The trade-off for banks is that the control of the customer interface belongs to someone else, just as when banks engage with customers on Facebook or Twitter.
It is easy to be impressed by chatbots, but a chatbot does not necessarily contain any intelligence. A scripted chatbot is just a collection of if-sentences, and will do absolutely nothing if presented a command that is not recogniced by the programs predefined conditions. (I even programmed a whole bunch of these for IRC myself back in the days. Each one was more useless than the other). A.I. chatbots on the other hand are based on natural language processing and machine learning. They area ble to learn and absorb information in a human-like fashion. Thus learning user behavior and preferences in order to give a personalized customer experience and everyday financial advice tailored to each individual customer from the comfort of your own smartphone.