5 AI Case Studies in Banking

Banks are using artificial intelligence (AI) to uncover new business opportunities, keep up with changing regulations and provide better customer service, among other uses. The organizations are employing machine learning (ML), automation, chatbots and more to improve their operations and service levels. The following case studies provide some examples of how the banking industry is applying the latest AI technologies:

1. Commonwealth Bank of Australia

Commonwealth Bank of Australia needed a way to efficiently analyze billions of transactions, looking at both structured and unstructured data to uncover previously unexplored opportunities with documents, text and imaging data, according to a case study.

The bank used H2O.ai’s Document AI product in a strategic partnership arrangement that enabled the bank to gain early access to the technology. In four months, the bank was able to process millions of documents a day on the AI platform.

Document AI also helped Commonwealth Bank of Australia onboard new customers more quickly while ensuring compliance with risk policies and regulation. The AI technology automatically extracts critical details, such as name, date of birth and address from account applicants’ passports, driver licenses and other identifying documents.

“This really is just the beginning of where we can start to embed and re-look and radically reimagine our day-to-day operations and make lives better for our colleagues and our customers,” says Sonal Surana, general manager, Commonwealth Bank of Australia.

Results

Invoices can be processed 10 times faster
Automated matching of invoices, purchase orders and receiving reports
Accuracies and automation of 50%-85% on varying document types

2. Banca Mediolanum

With shifting economic forces affecting customers’ income, cash flow and ability to make debt payments, Banca Mediolanum needed a way to quickly recognize customers’ credit issues, particularly as the regulator was changing its definition of default, according to a case study.

The bank worked with SAS and its Viya product to develop reliable credit technology around default, incorporating AI. The technology enabled the bank to smoothly adapt to the updated European Banking Authority definition of default as new regulations came into being, requiring banks to follow stricter default criteria and change the way that they classified clients.

Though initially chosen to help with the bank’s risk management, the technology has also been a catalyst for innovation for other business units.

“Banca Mediolanum has used machine learning and deep learning techniques well before the new regulations were put in place,” says Stefano Biondi, chief risk officer, Banca Mediolanum. “This allowed us to consolidate our use of advanced technologies in model validation.”

Results

Developed several new scoring models and verified their validity during implementation
Improved credit scoring accuracy and customer service
Paved the way for the development of new credit products

3. Federal Bank Limited

To serve its over 10 million customers, Federal Bank Limited needed an AI personal assistant that could provide natural responses to actual queries and be an essential element of Federal Bank’s digital transformation, according to a case study.

After a technology partner introduced Federal Bank officials to Dialogflow by Google, they saw possibilities to reimagine the customer experience through an interactive, auto-learning virtual agent that would be more than an off-the-shelf solution and allow the bank to easily add its applications.

“The questions can be put in any colloquial language and need not be phrased in a textbook manner. If the customer asks, 'How much money do I have?,' the AI will understand they are looking for their balance and provide the correct answer,” says Jithesh PV, head of digital, Federal Bank Limited.

Results

98% accuracy in answers to customer queries
Enabled virtual banking to scale by more than doubling handled queries to 1.4 million a year
Expected to save 50% in customer care costs through AI automation by 2025

See more: How AI Is Being Used for Consumer Education in Banking

4. NatWest Group

Shifting mortgage regulations and processes can complicate the home buying experience for consumers, so they often seek out information and clarification from lenders. But the ongoing industry changes can challenge even the most knowledgeable mortgage employee.

NatWest Group worked with IBM Consulting to build an AI-powered platform to help employees guide prospects through the home buying journey, according to a case study. With the solution, nicknamed “Marge,” NatWest employees received quick digital mortgage support from Marge by typing keywords into a console. With cognitive enterprise technology at their fingertips, they were able to better support new and existing home buying customers.

Marge is embedded in NatWest’s existing data structures, with new data added continuously from content updates and customer interactions.

“In the mortgage industry, change is constant,” says MaryAnn Fleming, head of home buying services, NatWest Group. “Regulation changes, products change, processes change. It’s imperative that the customer has the support and the information they need to allow them to focus on their home buying journey.”

Results

Customer loyalty improved 20%
Call duration decreased 10%
Increased agent confidence

5. Valley Bank

Valley Bank’s anti-money laundering team sought to slash the manual work used in predictive modeling, which not only is time-consuming, but also results in a high percentage of false positives – or flagged transactions that are not money laundering activity, according to a case study.

The bank worked with Data Robot’s and its AL Cloud solution to quickly build and validate predictive models that more accurately flag suspected money laundering transactions. The technology tests and deploys models with a minimal amount of effort from Valley Bank staff.

“Our team has to work through a lot of noise to find productive information, which results in staffing challenges and a challenge in keeping staff engaged,” says Jennifer Yager, director of financial crimes compliance, Valley Bank. “When so many are false positives, you run the risk of missing the items that provide actual value.”

Results