Dave Basson, Advisor to General Insurance & Broking, at NTT DATA UK&I
Whilst it’s certainly a hot topic across the industry, artificial intelligence (AI) has still largely remained more of a buzzword than a tangible tool for UK insurers. Of course, some insurers are piloting AI implementations, but widespread adoption remains limited. The UK Government Actuarial Department’s recent report revealed AI/ML use is currently limited among actuaries across the profession.
Why the disparity? The benefits for insurers should be clear. Improved risk assessment, reduced costs, and enhanced customer experiences are all results that can be achieved with AI. Insurers can no longer have any doubt. In the future, AI will be an indispensable part of their toolkit.
Despite this, the industry’s hesitation is clear. Firms have concerns, fuelled by growing uncertainties, over the risks associated with embedding AI-powered technologies into traditional systems. Insurers fear impending regulation, the incompatibility of legacy technology stacks, and the possibility of infringing on customer privacy, among other factors. Fanning those flames is the constant buzz around AI in the media, and it is high time for some simplicity and clarity.
Taking the plunge
Let’s dive into the factors behind the industry’s hesitation. Part of the issue at hand is that AI solutions are not necessarily built for enterprises. Whilst ChatGPT, Bard, and Claude, among many others, have been hugely successful and popular within enterprises, they have not been specifically designed to be embedded within an enterprise technology stack.
Let’s take a life insurance provider as an example. If it was to make the decision to implement an AI solution into its technology stack, the first crucial step would be to set up a governance team as well as an ethics and compliance input to sit behind its firewall. Currently, AI’s propensity for inaccuracy, bias, and unreliability is a blocker for insurers to fully onboard themselves. Whilst there are providers that have created enterprise versions of their solutions – ChatGPT Enterprise being the strongest example – a truly in-house solution seems to be the only way to negate the technology’s associated risks.
Looking at all of factors behind insurers reticence to take the plunge into AI, some of which I listed above, the common thread is a lack of compatibility. The answer then, is for AI to meet insurers where they are, quelling the concerns of insurers over risk, regulation, and privacy.
Beyond the buzzword
With so much noise around AI in insurance, it can be difficult to separate the wheat from the chaff and gather genuine use cases. If you’ve ever felt frustrated at the lack of clarity, don’t fear, you are not alone.
In reality, the implications of AI for the insurance industry do not have to be complex. Many people have understandable concerns over job losses or costly data leaks due to the often-apocalyptic tone of media coverage surrounding AI, but it’s worthwhile to map out use cases in a clear and concise way without the added drama.
As with any emerging technology, there are risks, but there are also real-world applications that can bring tangible benefits today.
NTT DATA is focusing on improving customer experience by integrating AI into customer service operations. The technology will identify potential vulnerabilities customers may have and notify our contact centre agents and brokers. This, in turn, allows them to adapt their communication style to be more empathetic and understanding.
Additionally, AI will help simplify the complex terminology commonly used in insurance. This will give customers a clearer understanding of the products and services they are purchasing. This is particularly pertinent with the FCA’s Consumer Duty in mind.
Lastly, AI will enable insurers to respond to customers quickly and accurately, especially in times of need. For example, if a customer is involved in a serious car accident, AI can identify the case as high priority. It will then connect that customer to an employee specially trained in crisis management and compassion. This presents a more human face to customers when they need it most.
When it comes to AI, it’s a case of emphasising the art of the possible. I’m confident that as soon as there is a series of successful case studies, industry hesitation will decrease, and widespread implementation will begin to take force. AI needs to be seen as an augmenter rather than a replacer of humans. Ultimately, it will help insurance professionals perform their jobs more effectively.
Tangible change
If I had to choose two key areas where AI will impact the insurance industry most, it would be productivity and complying with the Consumer Duty Act. We are already seeing a vast array of use cases to tackle pain points in both of these areas, and these will only continue to grow.
Ultimately, the most important next step is communicating the benefits of AI in a way that is easy, fast, and clearly applicable to their daily work. There are clear benefits to productivity, and the message needs to be that AI is not derailing the industry, but simply moving it to a different track.
