The insurance industry’s widened use of data and analytics positions it well for AI productivity gains, as many different insurers have been gradually leveraging AI technology for a number of years.
In a recent report from Moody’s Investors Service, analysts highlight how deeper integration could offer opportunities to boost operating efficiency, which could benefit insurers’ credit quality.
The latest evolution of AI has been the commercialization of genAI, a subset of artificial intelligence that can create humanlike content, it is through text messages, images or sound.
Many companies are still in the early days of discovering this technology and understanding how to use it. However, some themes have begun to emerge.
GenAI can fast-stream tasks such as information summarization, content creation, intelligent search and coding. Analysts stated that by combining the power of genAI with insurers’ in-house data, could result in providing a “seamless, natural language-based experience” for underwriters, which ultimately will allow them to quickly synthesize multiple sources of information and so enrich and accelerate their engagement with policyholders in coverage discussions.
Moody’s also explained how the potential for genAI to augment business productivity makes it an “attractive investment” for insurers.
Analysts suggested that only deep AI integration paired with proprietary data will create significant competitive advantage. Some insurers will use their inhouse data to train the models, incorporating their own experience, customer insights and risk appetite levels.
Moving forward, analysts stated that it can expect to see insurers increasingly use AI to improve their customer service.
There is very strong evidence that backs this, as insurers have been using chatbots for a considerable amount of time to provide customer assistance. A key example is from Allstate Corporation, who helps small businesses with coverage questions through its AI-powered chatbot “Ask ABIe”
These virtual assistants are able to provide customers with 24/7 support. They can also address frequent questions, reduce costs and response time, while learning about customer behavior patterns to offer personalized products.
In addition, analysts also pointed out how some insurers and brokers are leveraging the capabilities of AI to use client data to extrapolate which clients may be interested in certain products. These decisions are made by comparing the profile of the customer with other customers with a similar profile.
However, as this technology continues to grow, key issues regarding data quality and privacy are often remaining one of the main topics of conversation in regards to AI.
Analysts explained that for those insurers who use AI, must ensure there are no biases or discriminatory tendencies within their models.
Indirect discrimination, which is a significant issue linked with AI models across the insurance industry, often stems from proxy variables and opaque algorithms. This type of discrimination can surface in pricing structures.
Further, as technology is further embedded into daily activities, some roles within companies will also need to change and new roles will be created.
As a result, businesses will need to assess the skills of their current workforce and formulate plans to retrain and up-skill talent so as to deal with process and technology changes.
Another major issue with the technology surrounds how much it could wind up costing insurers. Both the development and maintenance of AI models can be costly, which could also result in vendor dependency, especially in the realm of genAI where a handful of providers own the best-performing models.
One other major risk within AI models is that performance tends to degrade over time as patterns and relationships evolve.
Analysts highlighted how insurers will establish processes and tools to automatically detect performance issues and retrain models, which can be both financially and organizationally challenging.
