Man vs. Machine: Is the future bot or banker?

To cope with this change, financial institutions are now increasingly embedding a new digital way of thinking. Let machines do what machines do best, which is compute vast amounts of data using complex mathematics. Then, use those machines to enable humans to do what humans do best, which is solve problems, collaborate, plan, design, lead and inspire.

AI is now common in many areas. Within financial services, technology ranges from the sophisticated, such as algorithm trading and credit scoring, to the more mundane, such as automated expenditure categorisation. However, there’s a lot of scepticism about its potential consequences. Critics typically predict it will spark redundancies in a variety of professions. For example, driverless cars wouldn’t just reduce fuel costs and the number of traffic accidents, they could also eliminate the livelihoods of around five million professional drivers across the EU-28.

While we must acknowledge the changes new technology will bring, we must also focus on the opportunities. What new markets and services will AI enable? In the driverless cars example, there’ll be a whole new industry built around the software and hardware that enables them, creating jobs for engineers. And former drivers might become crucial testers of AI systems, using their experience to improve the technology.

People hold a lot of advantages over AI. But machines are much better at analysing vast data sets. They don’t experience inefficient Mondays or lack of motivation. And they don’t forget complex mathematical formulas.

They are, however, subject to wrong rule sets and low-quality data – issues experienced humans would pick up on. All this implies that AI will not replace humans but provide efficient tools and digestible data to support us. If we define the right rules and provide the necessary amount of high-quality data, machines will continue to be a game-changer for high volume tasks and fact-based decision-making.

AI should support humans

Financial services should harness AI to replace tedious jobs, freeing their people for more valuable tasks. For example, Net Promoter Scores (NPS), the likelihood of a customer promoting the service, are important for training, and measuring and improving service levels. But it’s often difficult to persuade customers to spend time providing this feedback and measuring the NPS is monotonous. By combining machine learning and natural language processing, it’s possible to build an algorithm that analyses customers’ tone of voice during interactions with call centres. The system then uses the tone of voice, or sentiment, to predict a customer’s NPS score. We created such a service e for a Danish insurance company and managed to generate a predicted NPS score with an accuracy of up to 90 per cent. This meant the company could avoid asking customers annoying questions, and free their customer service staff to build better relationships.

Insurance claims can also benefit from more prevalent AI technology. Image recognition, for example, can automate the process of determining the value of a claim, streamlining what is a tedious and expensive process for humans. The customer would take some photos of the damage they’re claiming for, and the image recognition algorithm would assess the damage and estimate a cost. This would improve efficiency and release employees to focus on more complicated claims.

AI-powered recommendations, like those we see on Amazon, could also be invaluable in financial services. By analysing customer data, such as the liquidity of their account balances, AI could predict when the customer might need a loan. It could even group similar customers based on spending patterns to recommend products others have used. With Open Banking making more data available across Europe, and various open source data pools offering richer information on customers, such technologies could be incredibly powerful.

For instance, we helped a Dutch insurer predict their customers’ needs in real time using both financial data and open source information about the wider world. As the insurer had limited data on their customers, we combined it with data from news feeds and popular searches to create a dynamic big-data platform. We  then built powerful algorithms to map all the information and predict everyone’s top of mind issue, anticipating the customer’s next question before they asked it.

The future is a hybrid of man and machine

While conventional financial services aren’t going to become completely AI-based service providers overnight, change is happening. Productivity gains will come as firms empower people with powerful tools, automatically analysed data and back-office automation. And with AI, big data and machine learning advancing at an unprecedented pace, the transformation is happening much more quickly than most players would like.

To remain competitive, they need to evolve just as quickly the technology, starting with a decisive strategy on how to best leverage AI. In our experience, that strategy should start with more back-office automation, more AI-powered general processes, such as customer service and fraud prevention, and more AI in decision-making.

The strategy must also incorporate a view of the medium-term future. Financial services have an opportunity to use AI to reimagine processes and services. Where human and physical limitations once constrained financial services, AI will unlock new and creative auxiliary financial services, limited only by people’s imagination.