AI in Banking: Case Studies

. 4 min read

Banks play a very important role in the development of financial life of modern society. Thus it is vital that every financial transaction done through the banks be properly documented.

Possible applications of AI

AI in the banking industry can interact with humans by making decisions, getting insight into customer preferences, ensuring that the customers are happy with the services provided by the bank, and help the customers understand their expectations from the bank. Let's look at some concrete applications:

  • Anomaly detection can be used to increase the accuracy of credit card fraud detection and anti-money laundering.
  • Credit underwriting evaluates the risk of lending loans to customers.
  • Image/face recognition using real-time camera images and advanced AI techniques such as deep learning can be used at ATMs to detect and prevent frauds/crimes.

Benefits to banking

  • Easy cash transactions (debit and credit) in less time.
  • Banks provide financial assistance or loans to people in need.
  • Online banking is easy and effective way to manage money.

AI is strengthening the competitiveness of banks through:

  • Enhanced customer experience
  • Prediction of future outcomeS and trends
  • Cognitive process automation
  • Effective decision-making

CRMNEXT

Analysis

CRMNEXT is a 360-degree view, customer-centric application for banks that helps manage the entire customer lifecycle. Relationship managers, bank managers, and personal bank authorizers maintain the generated leads and service interactions for customers.

A indicative customer record in CRMNEXT

Data elements

Case study 1: Understanding customer sentiments

Problem Statement: Lots of information to be retrieved from CRM to know the customer and decide the next cross-sell pitch.

Objective: 1-2 pages of AI-based customer insights empowering RMs with brief insights across customer sentiments.

Summarized report of insights

Data

  • Service interaction captures first line customer interaction needs and provides a fully integrated single screen solution. RMs enter comments based on the mode of interaction with customer. Branch manager reviews the comments entered by the RM and gives feedback to customer by recommending products.
  • HNW is a process derived from service interaction in which the bank engages with customers to understand their needs on products and addresses any kind of queries or services that is required.

Methodology

Sentiment Analysis is a NLP technique that identifies and categorizes opinions expressed in a piece of text. It then determines whether the attitude towards a particular topic, product, etc. is positive, negative, or neutral. Natural Language Processing data standardization has following steps:

  • Remove tagging: Perform the part of speech tagging where in the grammatical words like noun, adjective, adverbs are tagged out to know the actual meaning of text or sentence.
  • Remove noise from the new dictionary by removing stop words which don’t convey any meaning.
  • Replace by dictionary: Update the dictionary post removal of the noise or unwanted text.
  • Regular expression: Perform a pattern matching mechanism finding the relevant sentiments from the comments.
  • Stemming and lemmatization: An abstract form, shared by word forms having the same stem, part of speech, and word sense, stands for the class of words with same stem.

Solution

  • There are comments entered by RMs, sales team, and branch managers as part of the service interaction.
  • Machine will understand sentiments from the text or comments, i.e. positive sentiments like 'already done', 'good fund', or 'well being already covered', and negative sentiments like 'not interested', 'contact no. not reachable', or 'not banking', etc. using natural language techniques.
  • Machine will be fed with the raw data for a period of time based on which it will learn and predict results.

Case study 2: Understanding customer behaviour

Problem statement: To determine how much can be spent to acquire or retain a customer and on which groups of customers you should concentrate the most to maximize returns.

Objective: Customer Lifetime Value (CLV) measures all the future profits a given customer will generate in total for banks. Knowledge of this value is crucial in maximizing business efficiency.

Data: Using Customer Profiler information, the bank can segment their customers and predict a personalized targeted product offerings, by grouping similar customers based on their household income, demographics and age group.

Methodology: Customer segmentation is the process of dividing customers into groups based on common characteristics like annual income, savings, gender, age etc., so that companies can market to each group effectively and appropriately.

Solution: Product recommendation solutions work as information filters that help to present products that are likely of interest to the customer. Such services assist banks to boost sales and expand the number of returning customers. Bank can segment their customers and predict a personalized targeted product offerings using a recommendation model.

Benefits:

  • Planning campaigns to address the most valuable customers
  • Achieve better customer alignment and increase the impact on their purchases
  • Take right step to maximize revenue per customer

Conclusion

With customer centricity being key, a banking CRM has become the modern banker’s need. It is nearly impossible for banks to process and understand this data manually without a digital tool helping them out. In the quest for a customer-focused business, banks have started adapting CRM’s data using analytics, AI models that help them to identify selling points that would not be possible otherwise.

Further reads:   Artificial Intelligence in Banking       Bank Marketing Analysis



Get Started - Future proof your career

Join 150,000 aspirants. Learn Today - Apply Today. Try Free Programs

Learn Data Science Free with GLabs