The enterprises in the 21st century have been drawn into the data endearing ecosystem that is currently present today. With the formula of
more data, more success gaining popularity amongst several decision makers, it is highly evident that enterprises are focussed on quality customer intelligence, which they can integrate into their systems. However, there are several cases where enterprises
over-expect on their ROI, or they work on an inefficient banking model which does not deliver results.
Now, these factors are not factors that can be taken for granted. These are challenges which enterprises face when they decide to introduce AI into their banking system. In the digital world, there would be several successful banking mechanisms. However,
for every successful mechanism, there are multiple bad banking models as well. Hence enterprises should take note of what makes a banking model stand out.
So now, what is AI? And what is the massive conversation about it? Is the conversation worth the hype? Simply put, AI is what allows Amazon to know which product you want to buy, while Uber knows which car would be best for your ride, while Google
warns you about your hectic calendar. At the same time, banks try to stop a fraudulent transaction from happening on your account. This happens due to the studying of data by your AI/ML implementation in the core application systems.
The Question that a Bank Needs to Ask Itself
According to MIT Sloan, the implementation of artificial intelligence with the right strategy can increase banking revenue by 30% and decrease costs by 25%.
Amongst the different industries spanning a citizen’s daily life, the financial and banking industry would rank among the highly ranked industries which take up the highest priority. However, with the digital initiatives that have started to creep up upon
multiple industries, payments are also going digital with several outsiders challenging the banking industry with several digital payments options and mobile wallets.
Therefore the question that needs to be asked is what is the risk of not pursuing an AI approach to your core banking system? What is the risk in continuing with the legacy systems in banks and missing out on a major advantage over your competitors?
Having a customer-centric approach would be the first step towards obtaining a successful banking model.
Being Customer Centric – Baby Steps in the Banking Model
According to the HBR Closing the Customer Experience Gap Report, 93% of enterprises consider customer experience to be the most important factor.
Since data is the ultimate ingredient that helps in producing valuable insights and moreover, the ability to detect patterns in customer behaviour, it is vital to have the customer at the centre of your entire business model. Following the customers’ activity
and being customer-centric has never mattered more for an enterprise than in today's technologically advanced sphere.
Restructure your Banking Idea through Personalized Recommendations
Being Customer Centric means to segment your customers based on their priorities and recommend personal suggestions for their taste group. Retailers and media and entertainment giants have used collaborative filtering and content-based filtering to segment
their customers effectively. This could be an aspect that could be integrated into the banking space. This includes having a user-driven filtering, item-item filtering and content-based filtering.
Personalized Recommendations are the highest forms of providing excellent customer services by segmenting your customers.
Item-Item Filtering can provide you with a recommendation of banking services based on transactions made on similar services by other customers at the bank.
User-Item Filtering can provide you with a recommendation of banking services based on transactions made by similar customers.
Content Filtering can provide you with a recommendation of banking services based on past transactions or customer preferences.
The 4 Pillars of a Superior Customer Intelligence Strategy
In Baseball, the batting team can obtain a home run by running across the 4 bases. The bowling team can get the opponents out by guarding the 4 bases thoroughly. Similarly, apply this strategy to your core banking system. If you cover each aspect of customer
intelligence, then where can you lose out – The answer is, you have lots to win and none to lose.
Pillar 1: Data Discovery
As an artificial intelligence framework is only good with the influx of data and where its accuracy is directly proportional to the amount of data present, it is important for banks and wealth management firms to focus on data discovery
as it sets the stage for the rest of the framework to take place.
So where does the data discovery phase start? Data discovery requires the gathering of requirements. There needs to be a complete gap analysis where decision makers and data scientists can understand the structure of the current core banking system
and the structure of the proposed banking system. It is also highly important to highlight the technology stack that can be used to effectively approach and enrich the banking system.
Pillar 2: Data Optimizer
Once the gap analysis is performed, it is highly necessary to bring forth your data extraction process. There are different stages which take place within the data optimizer stage.
Banks need to understand about the various customer touchpoints across which data can be extracted. Customer touchpoints can vary across different platforms. Some of these touchpoints could be from customer payments such as purchases through cards, payments,
invoices, statements or bills. Customer touchpoints could arise from complaints and feedback on the banking portal or on social media platforms where customer emotions can clearly be put forth. It could also come forth from responses to discounts, usage frequency
and engagement history through calls and SMS.
Once the data is obtained across different touchpoints, the data requires cleansing. The data that is extracted from different sources could be of structured, unstructured, video or audio or image format. This is a vital stage of your data optimizer stage
as data which is collected usually contains a huge amount of incorrectness. The percentage of duplication also could be high as several customer touchpoints are involved in this stage. There are also several outliers who drastically change the insights or
projections for your data. Hence data cleansing should be done.
Data Standardization and Governance
Once incorrectness is removed from your data lake, there needs to be standardization and matching process. The ultimate output of this step would be to obtain a record that identifies all customers based on their different touchpoints to create a unique
record of the customer. It also would be to understand and segment the customer more intently.
Pillar 3 - Customer 360°
With the identification of all customers, the master data management process provides the platform for a 360° birds’ eye view of the entire customer record. The truth is data is virtually everywhere and experts have claimed that the banking industry is among
the highly ranked industries in terms of quantity of data. Hence it is a case of finding out what you already have in your bucket. The Customer 360° understanding of your customers’ data enables you to obtain one true version for your data while acting as
the nerve centre and enabling the delivery of data to relevant stakeholders at the bank.
Pillar 4 - Customer Intelligence
As a part of covering the four bases to achieve desirous ROI values, it is important to gain insightful predictions and analysis for the future by feeding data into your machine learning algorithm. The self-learning ability of your machine learning algorithm
gives you the freedom to feed as much of data as possible in the view of it recognizing patterns amongst the several million customers that your bank would cater to.
Followed by insightful analysis and predictive analytics, it is necessary to gain suggestions and feedback on the future journey, so as to enrich your core banking system with accurate recommendation engines and intelligent search options, while segmenting
and categorizing your customers based on their actions. Successfully covering this base also allows banks to make key decisions on product offerings and cross-sell and up-sell campaigns.
Artificial Intelligence Tech Stack – What Could Enhance Your System?
The core banking system consists of front, middle and back-end operations and each division have the potential to increase productivity and efficiency. The degree of the change in efficiency depends on the value of the implementation of artificial intelligence
and where it is implemented (i.e. the AI strategy).
So what should be present inside a superior AI framework?
Robotic Process Automation (RPA)
According to Automation Edge, RPA will grow to become a 29 billion dollar industry by 2021 – arise from the 250 million dollar valuation that it received in 2016.
RPA is a highly up-and-coming area where banks are looking to trespass into. Currently, there are several monotonous tasks that are present in banks which increase resource hours and take a longer time to market. As a result, operational costs climb upwards
while the banks’ time to market statistic suffers. With the introduction of RPA, repetitive tasks get handled faster and releases can be taken to market quicker.
Creating an algorithm where data is fed into your system for insightful predictive analysis on the customer data regarding several aspects such as loyalty, satisfaction, retention, risk, profitability, credit risk, lifetime value and so on. The entire collection
of data once fed helps in giving the bank insights on how to proceed with that unique relationship with the customer. Machine learning, in the end, helps banks to provide better enriching services and experiences for their customers.
With a variety of machine learning models such as supervised learning, unsupervised learning, hybrid learning and reinforcement learning, a strong machine learning algorithm would be present, which on the basis of the data present, would be able to classify
customers for rewards and penalties, while at the same time suggesting the next best action for the customer.
Image, Speech and Text Recognition
The role of image, speech and text recognition comes when translating data across various platforms including social media activity and webpage activity. Across different sites, customer behaviour can be studied through the analysis of audio, video and text.
The recognition of text recognition also comes into play with the use of RPA. Customer transactions that are recorded in real time could be in the form of text or images. The entire content of data needs to be transferred from its source into an observation
or record. RPA gives the possibility of performing the action without a particular human resource.
NLP & Customer Conversational Flow
With the inclusion of AI chatbots and robo advisors in the banking industry, it is vital to have a Natural Language Processing (NLP) approach where customer queries and requirements are clearly understood by the customer service robot.
69% of customers would prefer chatbots to humans on general banking queries and customer onboarding processes. Hence it is extremely vital to have an AI strategy with the incorporation of NLP and Conversational flow.
With the banking industry at the crux of digital transformation, it is important and extremely advantageous to implement an AI framework into the core banking system to assist banking technologists and innovators in enriching customer experiences. The banking
industry is a truck full of terabytes of data on which AI can act upon. Hence banks need a clear strategy to implement AI technologies within their system. If banks fail to act, there are possibilities for challenger banks and newcomers to use the available
technologies and acquire a sizeable amount of the entire customer demand. With a 1 trillion dollar revenue opportunity by the end of the next decade, it is highly important that banks take full use of the opportunity and partner with Fintechs.