How Insurance Organizations Can Achieve Real Impact Leveraging AI-Driven Data Insights

Published On: October 19th, 20225 min read

Let’s say you lead an underwriting organization that wants to shift its focus from growth to profitability. There are certainly different ways to go about this – operational overhead reduction, loss prevention, subrogation. But suppose your organization wants to boost profitability by improving risk analysis practices. By taking on more favorable risks and fewer unfavorable risks, your company aims to improve loss ratios and increase profitability.

Now, let’s say your path to improve risk analysis practices involves building an AI/Machine Learning model that can leverage historical loss data and use interconnected user and telematics data to present an optimal price to the underwriter based on targeted loss ratios. This sounds promising, right? In theory, a well-designed model will equip your company’s employees with the data insights they need to make better, more profitable underwriting decisions.

Many companies across a range of industries have done something like this, and yet… the projected outcomes based on their AI/Machine Learning investment never materialized. The fact is, even the most impactful data insights will often fail to unlock desired outcomes if the companies themselves do not – at the same time – realign strategies, processes, and incentives accordingly.

 

Having Good AI-Driven Information Is Often Not Good Enough

When investments in AI/Machine Learning deliver lackluster results, the data science division often takes the brunt of the blame. But it’s important to remember that what moves the business is not better data or maths, but the behavioral change that is enabled by the data science. Therefore, if you focus only on the data science without considering the support needed for the behavioral change, your company is likely to fall short of the desired outcomes, regardless of how good your data is.

This principle applies across industries that are embracing AI and ML models to guide decision-making, but it’s especially crucial for insurance companies that are lagging in digital execution already.

Simply put, having good AI-driven data insights is often not good enough. The right data must be paired with the right strategy, the right user experience to present the data to decision makers, and the right changes to organizational processes and incentives.

  1. A Clear Measurement Strategy: In addition to your AI-driven data insights, your company needs to have a clear measurement strategy that baselines three parts of the system, while also allowing for the tracking of gradual waypoints along the road to actual impact. The baseline needs to define (1) the desired outcome, (2) the desired change of behavior related to the outcome, and also (3) the statistical performance of the AI algorithm. Progress can then be measured in relationship to these so your company can know if the technologies delivered are having a meaningful impact. If, for instance, your company is not seeing progress towards the desired outcome, your team can analyze metrics around both the desired change of behavior and the statistical performance of the AI algorithm in order to zero in on a solution.
    • A UX That Enables Good Decisions: Stakeholders – such as in our example with underwriters – need to be given access to a platform or tool that embodies user experience best practices and empowers the users to make good decisions based on the most relevant data insights. Otherwise, these insights – if not presented in a way that can be consumed and understood – may be ignored or underutilized.
    • A Willingness to Change Organizational Processes and Incentives: Your company also needs to align existing organizational processes and incentives with the new strategy. Old ways of doing business could impede decision-makers from taking advantage of the new data insights. In our underwriting example, organizational incentives may need to move toward loss ratios predicted by the models presented. This would in turn rely on the organization becoming comfortable with the AI’s predictivity of loss ratios, which in turn relies on intentional change management.

 

Service Blueprints & Data Strategy

One of the best ways to ensure that your insurance company has the right strategy, the right UX and the right change management plan in place is to develop a service blueprint that shows the entire service ecosystem of your business. Everything from training to incentive structures to platforms should be mapped out. These steps can then be analyzed, role by role, decision point by decision point.

Going back to our underwriting example, a service blueprint can help you ask questions like: What are the account analysts doing? What is the underwriter manager doing? Also: What are the technologies they rely on? How are they incentivized? How do they track leads? All of these things that factor into the writing of risk can show you a holistic view of your business, your technology, your people, and how these come together to enable growth or profit.

And by modeling all this out, you can begin to see potential problems that may need to be addressed in order to improve your defined outcomes. So in keeping with the underwriting example, you might find that underwriters may lack access to telematics data, wearable data, geospatial data or other third-party data sources. And that lack of broad data ingestion impacts the data model predictivity. Identifying such a problem will inform how you go about designing your AI/ML data model, how you shift your data strategy, and how you will package the new data so that your underwriters can consume it and use it to inform their decisions. 

Or maybe your current business process reinforces the underwriters to look at broker-provided price targets because they are incentivized to win the work and earn the premiums more than they are incentivized to write profitable risk. By creating a service blueprint, your company will be able to identify points where existing processes might hinder employees from properly utilizing the new data delivered by the AI model.

If you’d like to hear more about developing a holistic data strategy that will empower your employees to achieve ambitious new targets, we invite you to connect with our Finance and Insurance Services team.