Credit unions and many other organizations have long admired Uber for its signature friction-free app. In fact, the ride-sharing startup is often hailed as being something of a pacesetter in user experience circles. What fewer people talk about is Uber’s emerging command of machine learning, an artificial intelligence technology experiencing exponential growth.
While the company is using its machine learning platform in several different areas of its business, the application of Michelangelo (Uber’s moniker for the platform) to UberEATS is particularly noteworthy. The food-delivery arm of the business, UberEATS is relying on machine learning for, among other things, delivery time predictions.
Because Uber is known and loved for its pinpoint accuracy on ride-share arrival times, the startup naturally aims to “deliver” the same experience to in-home diners. And yet, food prep and distribution is an entirely different ballgame – one that requires the intelligence of machines.
Michelangelo relies on an immense amount of data to work. Things like time of day, location, historical data and average times are collected, analyzed and used to calculate delivery time. More importantly, the data is leveraged for learning. Michelangelo pays very close attention to the past to predict delivery times more accurately in the future.
As machine learning and other forms of artificial intelligence are deployed in mainstream services like ride-sharing, consumer expectations for seamless, hyper-personalized and predictive experiences are only going to increase. How will credit unions respond?
Machine Learning in the Movement
Alongside its credit union clients, CO-OP Financial Services is formulating an answer to that important question. The credit union leaders who partner with CO-OP on innovative payments solutions are actively engaged in the co-creation of democratized machine learning and artificial intelligence platforms that will ultimately benefit the entire movement.
One of the main ways credit unions stand to benefit from mastery of this technology is differentiation. According to Aite Group, even the largest financial institutions in the country appear to have a long way to go before they achieve success with machine learning in action. Aite’s study surveyed financial institutions of $30 billion in assets and above. Among them, 30 percent had no plans to deploy machine learning. Just 40 percent had machine learning in production. Credit unions and CO-OP, on the other hand, will soon enable the technology for the benefit of the nation’s cooperatives and their members.
Predictability is crucial to providing hyper-personalized banking services. Using data that flows in and out of the CO-OP ecosystem, our analysts are deriving rich insights that predict what’s ahead for the credit union members we collectively serve. Models that predict everything from member attrition to credit card delinquency help people-centric financial institutions intervene to keep their members and the cooperative financially healthy.
Predictive Models Transform Banking
The big idea of machine learning is the more data we take in, the more likely we’re able to identify and predict performance. The CO-OP ecosystem, encompassing nearly 30,000 ATMs, the second largest branch network in the U.S. and more than 4 billion payment transactions, generates a vast amount of data. When machine learning algorithms that teach and adjust over time are applied, the insights are real – and importantly, they’re actionable.
Take, for example, two predictive models we are using right now to help credit unions execute credit line increase (CLI) and delinquency prevention strategies.
Each model predicts, up to 3 months out, which eligible cardholders are most likely to utilize a credit line increase, as well as which are most likely to become delinquent. Presenting credit line increase promotions only to those identified by the model optimizes revenue for the credit union. Similarly, taking a proactive approach to managing only the most-probable delinquencies puts efficiency at entirely new levels.
In the very near future, credit unions will have the ability to apply machine learning technologies and predictive models like the ones above in exciting new ways. They’ll transform the member experience, develop innovative banking solutions and perhaps even discover new forms of revenue. It’s an exciting time to be a part of the credit union movement, and we look forward to reporting machine learning and artificial intelligence updates here on Insight Vault.
If you’re interested in learning more about the opportunities for credit unions to truly leverage their data in today’s digital age, please sign up to receive the next issue of THINK Review Magazine: The Data Strategy Issue. And be sure to join us for THINK 18 taking place in Chandler, AZ May 7 – 10 where experts will dig into this topic and much more.