In addition to data, the complexity of the tasks plays into the strengths of AI in a banking environment. This is because AI methods open up possible applications for automating processes. Processes where conventional, rule-based software systems reach their limits. The detection of fraud attempts is a good example. This is where legislators’ enthusiasm for organisation, technical progress and the creative potential of criminals in particular come together. Dealing with the pace of change with a set of rules is hopeless. Someone or something is always going to be one step faster, which results in manual work, supported by inadequate IT procedures. AI applications provide a way out of this situation. When appropriately trained, they allow banks to achieve a higher level of automation in fraud detection and a faster execution of processes.
Fraud detection is just one area where AI can show its strengths in a banking environment. Whether reviewing applications, marketing communications or customer service: using AI applications pays off quickly. Costs decrease, speed increases, quality improves. A wide range of opportunities opens up for banks. The skill lies not in finding a deployment scenario, but in choosing from the numerous options.
Despite these circumstances, implemented projects with significant AI components are still the exception. This is what our survey of banking executives reveals about the AI situation at their institutions. Four out of five say AI is still in its infancy at their institutions. This is not because they do not attach importance to the technologies. On the contrary: 96 percent agree with the statement that investing in AI will lead to competitive advantages in the medium term. Where then does this discrepancy between what is possible and what is being implemented come from?
When it comes to AI, many banks lack experience and the willingness to take risks. For conventional software projects, the experts fall back on established procedures and learned technologies. They bring their experience from numerous comparable projects to the table. When it comes to AI, however, all that is still missing. AI applications also have the added quirk that it is not foreseeable from the outset whether the existing data will be sufficient for the desired application. The early phase of an AI project involves a lot of experimentation. And experiments can fail. Willingness to take risks is as important for developing AI applications as mastering technologies. Those responsible need to learn to live with that. Only then can AI solutions come into their own on a large scale.
Banks meet all formal conditions for the successful implementation of AI technologies. The task now is to create the right framework in which AI ideas and projects can flourish. And that is not a question of technology, but of will. And of courage.
Read our latest AI report for banks and financial services.