The Impact of AI-Enabled Insurance Tools on Risk Management

Tim Hardcastle


April 21, 2023

Impact of AI-Enabled Insurance Tools

In 2021, analysts with McKinsey & Company listed artificial intelligence (AI) among five tech trends predicted to transform insurance over the next decade. Yet it is likely even the experts at McKinsey had no idea how far—and how fast—AI would mature.

On Nov. 30, 2022, just 14 months after the release of the McKinsey report, ChatGPT burst onto the scene. The powerful chatbot from OpenAI reached an astonishing one million users in its first five days. In the weeks and months since, ChatGPT has sparked endless debate across all industries and is challenging assumptions about what risk management might look like in the next decade.

Clearly, the era of highly sophisticated AI has arrived, and it will forever change the way all organizations operate. Its impact on insurance and risk management is likely to be deeply felt long before the 2020s have concluded.

Technology with Deep Roots

AI takes information, interprets it and reaches a conclusion based on a wide range of data sets. This act of embedding logic into software is not a new concept. Consider Microsoft Excel, foundational software that nearly every risk manager has used. At its core, Excel is built around a fundamental algorithm: the if-then statement.

Innovations like ChatGPT show how AI has moved beyond the linear logic of Excel and toward something far more responsive and consequential. The exponential growth of computer power (especially as seen in the rise of quantum computing)—combined with technologies like AI, machine learning and deep learning—now allows for systems to collect enormous amounts of structured and unstructured data, ingest it, curate it and draw conclusions with impressive speed and accuracy.

The advance of AI is akin to the evolution of autonomous vehicles. People talked about driverless cars for decades. The first prototypes date back to the 1980s. But it was not until Telsa introduced its semi-autonomous autopilot feature in 2014 that people began real-world experimentation with self-driving cars. Today, major automakers like Audi, BMW and Ford are investing in their own autonomous vehicles. In the same way, ChatGPT has given AI its breakthrough. The question risk managers must answer now is how to benefit while avoiding potential pitfalls.

Benefits of AI-Enabled Insurance Tools

AI technology can achieve tangible business results for insurers including the use of chatbots to improve customer service, and data supplied by drones and satellite imagery to enhance underwriting. For risk managers, the benefits of AI and predictive analytics are even more sophisticated and impactful to the bottom line both in terms of claims management and risk mitigation.

For example, Milliman, one of the world’s largest independent actuarial and consulting firms, uses AI to estimate when claims will happen and the potential impact on the businesses with which they work. Through the application of predictive analytics, the company works with risk managers to identify sleeper claims, assess the likelihood of litigation, improve organizational processes and address skill, training and management gaps that may contribute to potential claims.

Corporate risk managers today expect their insurance partners to provide them with contextual, real-time pricing as well as a thorough review on their situational risks. The biggest positive for working with insurers equipped with effective, proven AI is their ability to deliver this type of analysis in an exceptionally short time frame, allowing risk managers to have real-time and data-rich analysis of their risk portfolio and the related costs of potential claims.

AI Issues to Consider

For all its promise, AI is still experiencing growing pains. AI outputs are not always accurate. Compounding the problem, when increasingly complex and dynamic data sources are employed, it becomes increasingly complicated for humans to discern fact from fiction. For example, in January, researchers from two universities asked ChatGPT to generate research abstracts, then had scientists review both the ChatGPT fakes and the real abstracts. Those experts could only identify the ChatGPT fakes 68% of the time. What’s more, 14% of the time, scientists incorrectly identified a real abstract as a fake.

Another potential drawback of AI is the risk for implicit bias. One tiny error early on in an insured’s policy rating, for example, could affect them for years. Consequently, this presents potential long-term implications for the risk managers seeking to best insulate their organizations from the very risks AI is sometimes promoted as helping to avoid.

Ensuring Future Success with AI

Risk managers will always want to work with an insurance partner they can trust. At the same time, risk managers also want to work with an insurer with an eye on the future. To best protect their organizations, risk managers should be sure to partner with insurers that have built a tech stack designed to make them more flexible and agile. This is an area risk managers are going to need to better understand going forward. To ensure they do, these best practices should be employed when evaluating insurers claiming to offer AI-powered insurtech solutions:

  • Establish your goals. Determine the specific need you have for your organization. Do you want to partner with an insurer who uses AI to automate or improve specific tasks such as risk ingestion, data enrichment or data visibility? Then make sure the insurer has the technology that can actually do what you want it to do.
  • Look for insurance partners with automation and straight-through processing. These types of tools will help reduce or eliminate manual processing, helping to create frictionless experiences, minimizing opportunities for error and creating efficiencies that will benefit the risk manager’s organization.
  • Check that their tech works with your tech. Ask the insurer how their technology will integrate with your existing technology and insist they provide case studies or testimonials on prior successful integrations. Look for low- and no-code options often to provide greater adaptability. Ask about the checks and balances they use to create their algorithms. These types of questions can give you insight into how sophisticated an insurer’s AI tools are and how well their products will serve your needs and work with your existing systems.
  • Go fast, but be cautious. Keep in mind AI is still developing. Find an insurance partner who is using it in a way that benefits your business but be realistic in your expectations.


Looking Ahead

AI has grown so sophisticated that building a homegrown solution is likely out of reach for even the largest of carriers. That means carriers will need to partner with providers to take full advantage of AI. Risk managers need to take the time to learn more about the AI technology resources offered by both their insurer and the service providers that implement such technology. This will help them understand the impact of AI on their business and determine which tools can offer the most effective insights into risk assessments and risk interpretation. As the rapid adoption of ChatGPT demonstrates, the technology is poised to change the face of risk management faster than you might anticipate.

Tim Hardcastle is the CEO and co-founder of INSTANDA, a provider of no-code insurance software for insurance carriers, MGAs and brokers.