If apps are like icebergs, then Oval Money is a very big iceberg.
Above the surface users enjoy a friendly and intuitive interface, with easy-to-use forms, menus, dashboards, and graphs. Below the surface, Oval Money is a machine that learns, using artificial intelligence to make complex functions seem easy, helping to create the best outcomes for users.
Several personal finance management (PFM) apps use “artificial intelligence” to help users keep tabs on their finances. But not all AI is created equal.
Oval Money’s development and data science team are using technologies and techniques used by some of the world’s most popular platforms, such as Facebook and Instagram. The Oval Money team employs the python programming language, which is considered among the best for machine learning applications, and are active participants in the PyData community.

The data mining and analysis methods employed by Oval Money are based on well-tested tools available through Scikit Learn, but customized extensively to deal with the specific personal finance management applications. Data analytics are powered by Pandas, and visualized through Jupyter Notebook. This allows Oval Money to employ complex random forest regressions and precise classifications and clustering to bring order to your spending and savings habits.
What is most exciting is how machine learning makes it possible for user input to shape the performance of the app.
Oval Money is able to learn not only from each individual user, but also from the wider user community. There are a number of applications where AI can make a what seems like a simple function very powerful.
In one key application, Oval Money applies “record linkage” to recognize recurring payments. This means that Oval Money can understand what you are paying for on a regular basis and help you make savings decisions around those payments. For example, Oval Money can learn about your habit of paying for a latte at the safe cafe each week, and can suggest that that recurring purchase be used as an opportunity to save, perhaps by rounding up the payment amount and saving the spare change in your digital savings account. At the same time, Oval Money can help ensure your money isn’t wasted unnecessarily. Record linkage can also identify duplicate payments, and notify you if an error from a merchant may have charged you twice for one purchase. Importantly, machine learning allows the user’s input to train the record linkage algorithm.

In another application of machine learning, Oval Money has taken inspiration from the latest research about savings in behavioral economics. One of the major barriers to successful savings is the impact of broken habits. With effort, individuals can get into a habit of saving their spare change or putting money aside every so often, but if an obligation, emergency or temptation arises and the regular practice is interrupted, it can be very different for savers to begin the practice of saving again. To solve this problem, Oval Money’s technology seeks to make saving less “deterministic.” Rather than rely on very strict rules around saving, Oval Money is able to learn when to apply flexibility in a user’s savings by adapting to user habits. Oval Money can identify when a user might need to save a little bit less in a given period in order to meet certain financial obligations, and it can actually suggest ways to readjust saving and spending to help the user feel in control even when such unexpected obligations arise. Moreover, Oval Money can draw on the experiences across the whole group of active users to analyse the reasons why people tend to need to reduce their savings, and understand to which group a particular user most belongs, suggesting savings strategies that have worked best for that group of users.

Clearly, creating a great user experience is just the “tip of the iceberg.” Oval Money’s ability to help people save is about using technology as a tool to solve problems. The fast pace of development in artificial intelligence and machine learning is really exciting, and personal finance management could be one of the most important applications for these emerging technologies.