Elula’s AI Solutions Help Banks Improve Customer Retention

Elula is helping the financial services industry in Australia solve business problems such as customer churn rates using artificial intelligence as a service (AIaaS) on AWS. Elula has built several AIaaS solutions, including Sticky, which highlights how engaged customers are with a brand. Elula delivers its solutions using multiple AWS services including AWS Glue to prepare and load data for analytics, Amazon EMR to process large amounts of data efficiently, and Amazon Athena as an interactive query service.

Startup Elula provides artificial intelligence as a service (AIaaS) to financial services institutions (FSIs) in Australia and New Zealand. FSIs are using Elula solutions to make predictions and business decisions and to accelerate the use of artificial intelligence.

For banks, the cost of acquiring a new customer is four times higher than retaining an existing one, so they are highly focused on client retention and reducing churn rates. Many are turning to Elula and its AIaaS solutions, such as Sticky, to predict customer churn and to retain those customers through better customer experiences. Josh Shipman, cofounder of Elula, says, “It’s more important than ever for banks to retain their customers, and our AI solution enable banks to identify high-risk customers three months before they leave and to have personalized conversations at scale.

To deliver its AIaaS solutions, Elula looked for a cloud provider whose services could easily scale to meet the workflows of banks. What’s more, the provider’s cloud platform would have to deliver the reliability expected from the world’s leading financial institutions and come with open source tools for Elula to develop and enhance the platform at pace. To this end, Elula chose to work with Amazon Web Services (AWS). Says Shipman, “AWS has a continued strong presence in financial services and can provide the flexibility and reliability we require. We could successfully build a world-class machine learning platform on the AWS Cloud.”

By working with AWS, Elula ensures that FSIs can protect their data assets with enterprise-grade security and privacy controls. “My background is security and, from day one, it has been the most important area,” says Shipman. Elula has several security certifications, including ISO 27001, which recognizes the company’s ongoing commitment to keep customers’ data well-protected. FSIs, for example, have peace of mind knowing that data assets securely stored using Amazon Simple Storage Service (Amazon S3) are encrypted and can only be accessed by authorized personnel. Shipman says, “We go through regular security reviews with the AWS team to make sure that everything is set up correctly. This includes Well-Architected Reviews and regular meetings with their architects. With AWS, we pretty much operate as one team.”

Elula can have FSIs up and running with their AIaaS solution—delivering commercial outcomes—in as little as 8–12 weeks. In that time, Elula will have the Sticky product fully trained and operational, which includes engineering, cleansing, transforming, and centralizing the data. Elula builds client data exchanges using Amazon S3, with AWS Transfer for SFTP managing file transfers seamlessly. It also uses AWS Lambda and AWS Fargate for internal data processing.

Banks can invest millions of dollars building AI solutions in-house. However, Elula offers a business-ready AI service that enables banks to avoid that expense, quickly deliver outcomes, and focus efforts on their core competencies. Elula has built the core of its AIaaS solutions using AWS services such as AWS Glue, an extract, transform, and load (ETL) service; Amazon EMR, which automates engineering for millions of rows of data; and Amazon Athena, which completes data cleansing. Plus, with AWS Fargate and Amazon Elastic Container Service (Amazon ECS), hundreds of Docker containers for the service can be launched to run thousands of machine learning models overnight to deliver results the next morning. Shipman says, “The more our models train, the more robust our results are and the greater the return on the provided data.”

Whatever AIaaS solution Elula is building on AWS, the focus is always on customer outcomes. For instance, Elula’s explainable AI (XAI) proprietary technology provides transparency and the reasons behind why Sticky has arrived at a recommended action. It uses multiple machine learning models and translates them into business language, improves false positive rates, and provides the best possible action for each individual customer. All data is then integrated with operations teams and their existing workflow. XAI is computationally intensive and by using AWS, Elula is able to process data in sub-second speeds and optimally manage costs. Shipman says, “Our products and services are constantly evolving and developing to adapt to changing business and customer needs. XAI is almost nirvana for a bank, providing transparency, trust, and deep insight into their customers.”

Today, Elula offers its FSI customers one of Australia’s largest big data platforms with a data science model development pipeline. According to one of Elula’s customers, startups like Elula are nimbler in meeting its business needs than traditional partners. Most importantly, Elula’s AI product delivers actual financial returns 12–18 months faster compared to the time customers would have taken if they were to build an AI solution in-house.

Their success is attracting others who have seen the attention Elula is getting across the financial services industry and have heard how they are reducing churn rates. “If banks are doing well, churn rates are at 6–8 percent, but there are some banks out there with churn rates as high as 20 percent,” says Shipman. “Banks need to reduce churn rates by only 1–2 percent using Elula, and they can be retaining several hundred million dollars in revenue.”