Where should banks be applying AI?
Relative to other industries, banks have been slow to harness AI. However, there are many functions today that are performed by humans such as: analysing information, making decisions or recommendations on what action to take that can easily be automated. Machines have the ability to absorb information from multiple sources and make decisions much faster than their human counterparts.
Functions such as research have always relied on human analysts to pore over company information and use experience and intuition to make recommendations. This is an obvious candidate for an intelligent machine and will allow investment research firms to cover many more companies and respond quickly to new information as it arrives.
Investment banking has many roles that have withstood the advancement of technology. Intelligent machines will be able to determine the optimal time and means for companies to raise capital; understand demand and select the right time for an initial public offering or when to issue debt and in which form. They will do this by pulling data from multiple sources including past issuance events and information from secondary markets far faster and more comprehensively than a human analyst could.
In the global markets, AI will breed a new generation of more sophisticated algorithmic trading, which is not simply about speed and hard-wired rules. In addition, similar to research recommendations, AI will also help to generate sales ideas and automatically disseminate these to customers at the most opportune moment. As banks develop their digital storefront, they will increasingly be delivered direct to clients over electronic channels. These could be, for example, an API or, increasingly, via chat, where the client is communicating with a bot rather than a human.
Whilst the most visible change will come in customer facing functions, it is likely that the first steps towards adopting AI will come from internal functions. There are many obvious applications in these areas, for example interpreting regulation and automatically translating these into rules that can then be used by machines to ensure compliance.
Other areas include operations, where many people that follow rules based workflow to ensure that trades are settled. Also, customer service functions will benefit from intelligent machines – which is already a target of many firms across many industries – to provide much improved customer support when there is a machine at the other end of the interaction.