A Platform Approach to “Using AI in Customer Experience” for Banking
By Jaymalya Palit, SVP & Head of Product
When banks and financial services institutions (BFSIs) think of providing a frictionless customer experience, they should start with the customer at the centre and think of everything else like channels, products and all other aspects around the customer.
This also means that instead of broad brush painting of nameless, faceless mass of customer groups, the BFSIs need to go beyond and look at individually identifiable segments of one.
BFSIs have databases which have a huge amount of customer data and if you add to this all the data available on social applications, it can provide them with enough insight to deliver individualized experiences to their customers.
The whole segment of one experience should be guided by the philosophy of “genuinely helping” the customer and not by revenue as the principal driver, following what Gartner calls “empathic banking”. Very rarely or almost never do we see BFSIs proactively helping clients at the time of need, except for maybe sending payment reminders. What about sending those payment reminders and also letting the customer know whether current and projected balances are enough to cover the payments as a start? Maybe even going beyond that to actually providing solutions in case payments cannot be made within the due dates?
A personalized customer focus like that, will build loyalty, advocacy and finally revenue will also follow. Will a customer not be open to listening to your cross-sell pitch after you have taken the due utility bill payment worry off his/her head? In this case the cross-sell can very well be just subscribing to your budget management solution which gradually takes him/her through the process of managing funds better and saving enough to buy the next insurance product you are distributing.
What can achieve this for BFSIs at scale? They have the customers and they have the customer profile and transactional data. They also have the reach through their own set of self-help channels like the internet and mobile and assisted channels like the branch and the call centre. So clearly the base is available and strong enough to move to the next orbit of differentiated and hyper-personalized customer experience. We will discuss in this paper a potential strategy to build on that base and move to the next orbit.
We believe banks must take a comprehensive and holistic approach of a customer experience platform and not use siloed, piece-meal, tactical design and technologies.
The first thing to look at is the type of interaction that can be the starting point of this experience. Structured form based menu options give the banks ways to capture transactional data but conversational, unstructured, natural language interaction is the best way to get closer to the customer, because only through conversations do people open up and tell us their preferences, opinions, feelings and inclinations. Even for simple queries, requests and transactions the customers need to remember the menu options and navigate to the right place and key in the right details to get their requirement completed whereas a simple instruction like “transfer 2k to my daughter” on a conversational channel can make the life of a busy mother that much easier when her daughter has run out of her pocket money in the middle of the month. She is already using all kinds of messaging applications for every other collaboration, so why not think of this interaction as a collaboration with the bank!
Think of another request like an amortization schedule or a tax statement which are typically available through 2–3 layers of re-direction starting with the “customer service” link and how this can be achieved through a sentence or maybe even just a couple of words. To top of all these simplification by the conversational interaction is the ability to understand and deeply engage the customer.
At the end of the month when the customer clicks on “balance inquiry”, mobile and internet banking can definitely show the balance in a blink but in a conversation when it comes as instead of “show my balance” to “how much do I have in the bank” one can detect the concern that probably too much has been spent last month and he does not have enough to cover all the bills he usually pays around that time. The more interactions the customers have in the conversational channel, the deeper understanding of the sentiment and mood at that point of time can be obtained. That can help the bank to provide the most appropriate assistance. For example when the query changed to “how much do I have in the bank”, the reply can be the balance of all the deposit accounts, the loan, card and bill payments due and maybe the projected balance in the next few days. What exactly is to be shown to that particular customer at that point of time should not be defined by the bank but should be completely derived by the system based on the learning of the conversation pattern and all the other possible information that the bank has collated over time.
The conversational channel can in fact start off at the prospecting stages as well for non-customers who are looking at the bank’s offerings. In the myriad of information, the prospective customer is often totally lost and needs help in deciding on the product purchase. A conversation experience that can provide the precise answer to their query can just prevent her from abandoning the request for a new credit card.
From the mode of interaction, we move to the channels. The examples discussed above are all in the cases of direct channels. So conversational channel can be added to the bank’s website for prospects and internet and mobile banking for their customers. The conversation channel can be extended beyond the banks’ own channels to where the customers frequent and not wait for them to come to the banks’ channels only. So a Facebook Messenger, Whatsapp, WeChat, Line, Viber and so on can all become channels of conversational banking. For a very quick interaction while listening to one’s favourite music, “pay my electricity bill” could be a real help through a voice channel like Alexa or Google Assistant.
Even for assisted channels like the call centre and the relationship manager (RM), the bank representative at the other end struggles to find the right option to reply to the customer query. Here also a conversation channel combined with the existing systems could make the agents life a lot simpler and help turn-around the customer query much faster.
An AI based customer exngagement platform that can cover all channels, be it messenger or voice, be it a direct customer channel or an assisted one, can power the bank to provide the deep engagement expected in this digital era. Only a platform approach can provide the scale and the ability to collate all the institutional knowledge to provide the real empathic banking. The experience has to be seamless, uniform and consistent across channels and different points of time. The customer centric journey we are talking about is all around situations, sentiments and emotions and every time it has to be contextual. A siloed approach of a chatbot here and a voice channel there can only provide a fragmented experience. An end to end AI platform can take the engagement several levels deeper through:
· Powerful natural language understanding (NLU) for the customer speak — starting from simple micro conversations to long, multi-level, complex instructions
· Continuous machine learning (ML) from conversations, reply appropriately and help out in different possible ways based on all the data and analytics collated from various sources. ML can engage the customers throughout the interaction with the next best step for that point of time and can provide one or more options for the particular context
· Learning from conversations in the assisted channels and use those learnings in direct channels and also help in assisted channels by learning the best practices from various experienced relationship managers and agents and prompt the new ones to use similar ways to help out their clients
· Direct and assisted channels and beyond the bank’s own channels — chatbots on internet, mobile and social channels, voice, call centres, relationship managers and even other unstructured natural language channel like email
We are seeing a mass adoption of unstructured natural language interaction by all of us with all the applications and systems that we use on a regular basis. BFSIs are already providing or thinking of providing a similar customer experience. Most of such initiatives are for one or two specific channels and are often using different technologies.
A platform approach to using AI in customer experience can provide hyper-personalization, a deeper engagement and a consistent and uniform experience through direct and assisted channels and even beyond the banks’ own channels. Irrespective of channels, the underlying AI stack should be able to understand conversations, learn from conversations, reply appropriately and help out in different possible ways based on all the data and analytics collated from various sources that already exist in the banks. True empathic banking at scale can be achieved only through this kind of a platform approach.
To know more on how to transform banking experience, please visit Active.Ai.