One of the reasons why generative AI is creating such excitement is that it promises to reinvent businesses across all industries, financial services included. But banks and insurance companies have, over many decades and in many areas, remained stubbornly labor intensive and inefficient. Given the cost and risk of transformation, together with the inherent conservatism of these sectors and the prudential effect of regulation, gradual change has always the more appealing option.
This time, however, they may not have a choice.
This is because generative AI is different than other tech innovations. Even the internet and mobile phones had limited impact compared to generative AI, which seems likely to affect every part of every financial services provider.
I am utterly convinced that the advent of this technology will prove to be a watershed for financial services, as it will for most other industries. But I also understand the skepticism and the caution. The question is: will this hesitance prove costly, and if so, how should the industry act?
The good news is that everyone is already working with generative AI. I think it’s fair to say they fall into three main groups:
- The first is doing just enough to persuade stakeholders that they’re taking things seriously. They have their toes in the water and they’re watching everyone else to see how, where and how quickly they’re adopting the technology. As with musical chairs, they’re doing just enough to avoid being caught without somewhere to sit when the music stops.
- The second group is more committed. They have identified the low-hanging fruit and are running proofs of concept so they can start to benefit from the innovation. While there’s a lot to commend about this approach—which is often influenced by a shortage of skills and investment funding, concerns about risk and regulation, and caution among their leadership—the failure of these standalone projects to reach beyond operational silos is likely to prevent them from revealing the full potential of generative AI and materially impacting firms’ bottom line.
- The third group—and I would argue that very few banks and insurers fall into this camp—has recognized this transformative potential. These firms understand that the technology is a game-changer, a terrific opportunity that can only be fully realized with a strategic, holistic approach. This includes fostering a culture of experimentation, embedding the right skills in the right places, and scaling adoption across the organization.
To illustrate this, think of a bank that looks to improve its customer service by introducing an AI-powered chatbot. Because its core systems have not yet been moved to the cloud, because much of its data is trapped within functional silos, and because its people have not been reskilled to use it effectively, it’s quite likely that the chatbot will disappoint. This will dampen enthusiasm, make it more difficult to gain support for other use cases, and empower the skeptics.
Banks in our third group realize this. While many may be harvesting their low-hanging fruit, they are simultaneously examining their digital core, their data foundation and their skills base—to mention just three key elements—to understand the changes they need to make across the enterprise to reinvent themselves and capture the full potential of AI. They also understand that a mastery of what constitutes ‘responsible AI’ is a critical enabler that accelerates adoption.
What differentiates this third group—in Accenture’s recent study on generative AI we call them the Reinventors—is their vision and ambition. Because they look beyond experimentation and individual use cases and instead envisage a transformed industry, they approach the challenge more holistically. They have a clear picture of what they want to be, and the services they intend to offer. This allows them to thoroughly tackle the obstacles, fundamentally reinventing their value chains and preparing the organization for a future that may be a moving target but will be very different than today.
One example is National Australia Bank (NAB), the country’s second largest bank. NAB established a specialized generative AI team to support business leaders with the specialist expertise and technology required to accelerate responsible experimentation, scaling and value realization. The aim was to rapidly test the efficacy and ROI of the use of generative AI in the context of solving real business problems. The bank established a production-level generative AI platform with banking-grade security and risk controls, and carefully tied its generative AI strategy to its overall business strategy. NAB is known as ‘the relationship bank’, so it has prioritized use cases that enable its people to deliver enhanced customer service while also improving their productivity. These use cases include an assistance tool that enables bankers to quickly answer customers’ queries and address their requests.
It’s easy to understand why most banks and insurance companies are daunted by the challenge of scaled, holistic transformation and why those in other, less-regulated industries, like software and platforms, life sciences and retail, are moving faster to fully embrace this technology. However, with generative AI the gap between the early and later adopters is likely to be wide—and will expand steadily due to the ability of AI to learn from its experiences and become ever more potent.
I believe it’s hard to overstate the potential of generative AI to reinvent our industry. I’m convinced it will enable organizations to renovate their digital core, transform their operating model and streamline their data architecture—and in the process enable the creation of innovative new services such as personalized products and advice. It will revolutionize how work is done and how customers interact with their banks and insurers. It will dramatically change the economics of financial services.
If I’m right, generative AI will be a breakthrough for financial services. But not for all, because hesitance to embrace it fully may prove to be a decisive competitive disadvantage.