Artificial intelligence in financial services: myth and reality

New technologies should be a means, not an end

“One of the biggest challenges is that people are looking for solutions without identifying the problems,” according to Matt Davey, Head of Business Solutions at Societe Generale. In other words, they should not just employ the technology for the sake of it, but have a vision of their objectives and then a plan of the most expedient way of achieving these goals. “Firms can lose sight of their business requirements and how best to use the technology” he says. “The result is that expectations may not be met but there certainly has been a lot of interest created. This is the case with any new technology.”

Where is AI when it comes to financial services?

The difference with AI is that it is not an amorphous technology but a collection of tools that are quite distinct and tackle different business issues. They are currently at various stages of development, but to date chatbots, or virtual assistants, better known to many as Siri and Ok Google, are among the most advanced. However, machine learning, natural learning processing (NLP) and robotic processing automation (RPA) are also gathering momentum. Together, they can generate huge efficiency and quality gains through validation of the data, providing proactive notification and realisation of patterns, identifying errors and producing trade and transaction reports for compliance purposes.

Natural language generation and understanding are also being added to the mix to absorb raw information in large datasets, understand and detect trends and correlations as well as identify inferences, risks or investment opportunities. Looking further down the line, NLP and generation will make it increasingly difficult for customers to determine whether they are talking to a human or an AI interface. Voice and facial recognition will also be further developed to improve the customer experience and for cybersecurity purposes.

RPA is also attracting a great deal of attention because it aims to replace the manual handling of repetitive and high-volume tasks. It differs from traditional automation software in that it does not have to be fundamentally redesigned and transformed. “While there has been a lot of focus on how AI can enhance the investment decision-making process, RPA is more about improving operational processes,” says Davey. “It is gaining traction in financial services because its scalability and efficiency make it a compelling business case. Banks have spent time and money on integrating legacy and new systems but you still need a human to rekey information. RPA enables you to parameterise the process and run it 24/7.”

Banks have spent time and money on integrating legacy and new systems but you still need a human to rekey information.  Matt Davey, Head of Business Solutions at Societe Generale

Aside from the benefits in integration, Davey points to several other viable business use cases in trade processing, reconciliation and client controls. “For example, RPA can also be used in setting up funds where people have to key in a lot of different details. The technology can define the process, automate it and then automatically make the updates.”

An adaptable legislation encouraging innovation

Not surprisingly, there could be a host of regulatory issues surrounding these new technologies, but legislators are treading carefully for now. For instance, in the UK, the Financial Conduct Authority is looking to roll out its so called “sandbox”, which it launched in 2016, on the global stage. The initiative allows financial service firms to formulate new ideas and develop products in a ‘safe’ environment plus offers support in identifying appropriate consumer protection safeguards that may be built into new products and services.