In the world of artificial intelligence, many shifts have taken place to allow for a more human centric approach to business. One such shift is financial wellness. This article introduces the concept of financial wellness and the practical implementation of using machine learning and AI to give customers a financial wellness score. There are further considerations regarding the concept, including why, despite it being an extremely relevant topic, it has not yet been implemented globally.
Financial wellness is defined in a number of ways; the most familiar of which refers to an individual’s overall financial health. It is often described as a position of ‘being prepared for an emergency’, ‘meeting financial goals’ and ‘budgeting spends’, amongst others.
The case for financial wellness in business is a problem that many big companies are currently attempting to solve. Some remedies include taking well-known concepts and altering elements to change focus, while others involve the capabilities afforded by machine learning. Because wellness is not a static construct, it requires a more nuanced approach to measuring. There is a need for frequent experimentation, model redeployment and constantly updated scoring. Alongside this, there needs to be a space in which customer interactions can be considered, to reveal whether contexts are changing. This dynamism is imperative to stretching analysis beyond the traditional statistical approaches to client behaviour. History is quantitative and fixed, the future is qualitative and limitless.
From a purely business perspective, understanding wellness comes down to behavioural and technical elements. Measured, tracked, tested and measured again. However, of vital importance to the human-centric approach to financial wellness, assistance directly to the customer is imperative. An interesting dilemma regarding the human/money relationship is akin to blissful ignorance, and this condition calls for careful consideration and gentle testing in order to overcome common behavioural barriers.
Financial Wellness poses a multidimensional problem, for which there are a number of proposed approaches to solving. Factors must be in place that allows for the altering states of the human condition to be identified. This makes for a very complicated machine learning and AI landscape. While typically approached in a quantitative manner, there are many qualitative properties to consider when using an entirely automated solution. Principles of changing behaviours can be derived from historical data. It is possible to ascertain these patterns based on constructs already in existence within the field of Computational Social Science. The technological understanding of shifting behaviours becomes vital to establishing trends of financial wellness, or lack thereof.
Without delving too much into the multitudinous nature of Computational social science: in a nutshell, CSS refers to a discipline which uses computational and automated analysis approaches to social science. The CSS paradigm provides the framework from which the technological approach to understanding humans, involves human knowledge too. Wellness is directly related to this by means of looking at it from the perspective of a company that wants to help people with their financial wellness. There are a few frameworks currently in place that fintechs use to help improve the human and social element, enabling customers to complete activities that will lead to financial wellness. Such as setting up a budget, having budgeting goals, saving, or paying off debts. These frameworks provide an idea of how to begin predicting where they’re headed in their financial lives through scoring mechanisms. No savings give negative scores, while a retirement plan gives positive scores. Based on these scores and social contextual information – such as peer trends, financial institutions can offer suggestions for score improvement.
The ecosystem.Ai approach to this process involves a continuous cycle of learning and updating customer profiles at the exact moment changes happen. A cold start intervention begins the process of testing, and as more information is gathered, its fed back into the system in real-time, and used to provide useful motivations for the customer to continue to improve. One approach to encouraging a drive towards wellness includes offering a view of a customer’s ideal self, and further offering solutions and steps that can be taken towards becoming that self.
Using the ecosystem.Ai workbench, it’s possible to productionize models. Selecting a model type such as Auto-ML, and linking it to your feature store, it will then generate the default hyper-parameters automatically for you when you do the generation. From there, putting that model into a production environment, you could assign an algorithm such as XG boost to use and test. The model will score in real time in a production environment, and then measure client feedback. It then continuously adapts to new parameters, updating features based on feedback changes, in the quickest possible time frame. Another function in the workbench allows for a continuous pipeline of the data science process. Using the ecosystem.Ai workbench, its possible to experiment with various hyper-parameters, and perform tests entirely independent of the production environment. Meaning that while you have models in the real life environment, you can also be testing other models for the same project without disrupting what’s already in place.
One of the best financial wellness breakdowns involves creating four pillars upon which to balance a customer’s financial health. The areas of focus are Spend, Save, Borrow and Plan. The ecosystem.Ai approach is to bypass traditional survey questions, which provide a view of customer stability from their perspective, and attempt to avoid the human bias using transactional data. This can be achieved a number of ways depending on the institution, and then combined with common pillar measures, which provides an overall financial wellness measure. By not relying on the customer’s unintentionally biased feedback, and automating the process, a truer representation of that customer’s financial situation becomes clear.
With so much information pertaining to the benefits of ensuring wellness in customers, why then, has it not already been implemented across the board? The most problematic element in the world of wellness is a difficulty in quantifying lost opportunities. When choosing to assist customers in doing, not what is good for the company, but what is good for them. This new practice changes the current ‘instant profit’ view of doing business, and turns it into a matter of lifetime value. Putting an emphasis on making sure that long-term engagement shows a mutual loyalty between customer and company, with a promise to serve each customer the right way for them. The correlation between customer lifetime value and company profit is still under investigation by a number of big companies, of which some have seen a noticeable uplift in general profitability, from moving away from exploitation practices alone.
Financial wellness is a topic of our time. With the capabilities afforded by machine learning, the resolution of which can lead to solutions for other business and customer problems. The current trends in favour of customer-centric approaches to business, mean that artificial intelligence has taken precedence. Using ecosystem.Ai products allow for the utilisation of machine learning functionalities without the complicated use of multiple disparate products. A single interface from which to build out the features needed to create feature stores, even if there are many of them in constant use. As well as constructing models to productionize, deploy and test using various feature store and algorithm combinations. The hypothesis relies on companies having the ability to continuously experiment and utilize models that fit their wellness constructs. Customer financial wellness is achievable, using machine learning and behavioural predictions. It’s now up to companies to make the investment in their customers’ financial wellness.