The global consultancy firm, Accenture, predicts that the growth of artificial intelligence (AI) and its impact on global industries will be exponential. The insurance and financial services market is no exception. In fact, Accenture recently predicted that the insurance industry is among those set to benefit the most, with growth forecasts of up to 4.3 percent by 2035. However, challenges remain—particularly where data is concerned.
Large data sets will fuel this growth by providing inputs for AI, according to MIT Sloan Management Review. The volume of data available to insurance companies means they are among the best placed to take advantage of AI.
Innovative players such as Allianz, which insures some 85 million customers in markets that include France, Germany, Italy and the U.S., is reducing premiums for drivers who use telematics to prove that they drive safely.
Similarly, American health insurer, Aetna, has entered a partnership with Apple to offer Aetna’s workers a free Apple Watch when they agree to track their health data, which may then inform their health insurance policy. AI can use this kind of data as inputs into machine learning systems and drive market insights and organizational agility and efficiency.
Despite these innovative efforts, some insurers are failing to make the most efficient use of that data. In the underwriting process, for example, many insurers are looking for ways to incorporate AI to automate. However, too often a major stumbling block is that much of the relevant data, both structured and unstructured, is stored in too many silos.
And this is a problem because, “Surprisingly, despite our world being quite literally deluged by data … a good chunk of it is not labeled or structured, meaning that for most current forms of supervised learning, it’s unusable,” says TechCrunch.
Despite this, insurers are asked about the potential benefits in their industry. For example, underwriting is widely front of mind as one area where AI could provide a significant opportunity to optimize processes and drive down loss and expense ratios.
Claims management is another area where AI could deliver value. Forward-thinkers in the industry have discussed how amalgamating data could aid case estimation. By combining historical in-house claims data with digitized, structured information, specifics of the claim being reviewed can be compared to other claims in the archive to make comparisons and decisions about the claim under review. AI could then be used to determine both selected claims development patterns and initial expected ultimate loss ratios without human input, or with very limited input, depending on preferences.
In a similar vein, insurers are looking at AI as a means to remove many of the arduous, manual processes they rely on today. For example, reserving for financial reporting or internal management information purposes is widely felt to be a significant, often manual burden on actuarial reserving functions. Many organizations currently segment their business into a high number of portfolios for reserving purposes, creating an onerous workload for reserving teams. Automating this process through AI will save time and money.
However, the barriers to entry for AI in insurance are both technological and cultural. Complex systems left over as legacies of previous acquisitions is the biggest barrier for a lot of insurers. A vast web of different systems and software creates its own intricate, disparate data landscape, which is difficult for insurers to untangle. To add to the problem, the technology environment is ever evolving, and for many insurers it is difficult to keep up.
Cultural challenges are part and parcel of an industry that has been used to handling data in a certain way. For the insurance industry, there is an inherent skepticism and fatigue following a history of failed data projects. It therefore makes sense that insurers now feel they must know exactly what outcome to expect from any proof of concept (PoC) before the project starts. This conservatism means many projects never even get off the ground, and it leads to thinking that any PoC will require years to solve a problem, when it might actually be doable in months. This is a double-edged sword. PoCs cost time and money, and they might fail—but they may lead to concrete results and a much-needed change to the status quo.
In spite of these challenges, conversations with insurance industry leaders suggest that there is agreement that this cultural barrier must shift, and that it can only happen through successful, fast track PoCs that deliver business value and are not science projects.
Ultimately, AI can help the insurance industry automate the entire process of information review, risk management and pricing, and enable better decision making. Its incorporation should be encouraged if insurers are to ride the wave of growth that is surely coming.
Like what you just read, here are a few more articles for you to check out or you can visit our blog overview page to see more.
Learn about data bias in AI, ways technology can help overcome it, why AI still needs humans, and how you can achieve transparency.
Successfully responding to changes in the business landscape requires data agility. Learn what visionary organizations have done, and how you can start your journey.
Sharing data can be relatively easy. Sharing our specialized knowledge about data is harder – and current approaches don’t scale.
Don’t waste time stitching together components. MarkLogic combines the power of a multi-model database, search, and semantic AI technology in a single platform with mastering, metadata management, government-grade security and more.Request a Demo