Originally published in Forbes, read it here.

‍

Insurers are investing heavily in building data lakes, increasing cloud scalability, improving data infrastructure and generating high-impact insights. These investments are particularly valuable when they're applied to insurers' executive decision making and strategy. Integrating existing and new data sources into decision making means a more informed or even automated process based on reliable and validated data rather than gut feel or intuition. Based solely on this, automated decision making that's fueled by data and insights instead of human expertise appears to be the pinnacle of decision making, but it's not — quite.

Automated decision making might be optimized for data, but that doesn't make it complete.

Why not? Why is a data-based, automated process an incomplete approach to decision making in insurance? There are a number of potential pitfalls:

1. Data is, by nature, historic, and the world changes constantly.

2. Data isn't perfect. This doesn't mean it shouldn't be leveraged or integrated into decision making; it just means it can't work unsupervised.

3. Data cannot replace industry experience and expertise.

4. Data doesn't care about the "why," but people do — including regulators, customers and shareholders.

5. Data can be biased and must be carefully monitored in case these slip through into the insurer's decisions and outcomes.

Automated decision making may have the data edge on traditional, gut-feel decision making, but there's a sweet spot between them. It comes from having a process that starts with all relevant, real-time data and ends with decision makers understanding the full scope of their options, relying on a combination of experience and data to make a choice.

Insurers looking to increase return on capital, decrease expenses and engage policyholders should consider that all of these objectives are attainable with the right information and decision making.

As data plays a more significant role in decision making, human perspective is needed for balance.

Reliable data is essential to good decision making. There are too many variables impacting insurers, shareholders and policyholders for it not to be incorporated as thoroughly as possible, especially when insurers now have access to tools and platforms that enable it.

Data is vital, but when important decisions need to be made — particularly those that impact the policyholder — the individuals in the room matter. Their expertise and experience with customers, partners, competitors and the market matters, and it should ultimately impact decision making regardless of the automated insights and scenarios new technologies can provide.

The way to ensure this is to get clear, useful data into the hands of decision makers who typically might only receive brief summaries or recommendations. These individuals should have the tools to understand the impact and trade-offs of different decisions as well as tools to monitor the outcomes of their decisions in order to adjust and refine the process moving forward.

How to properly balance data insights with human experience and expertise.

To begin striking this balance, insurers can start by identifying gaps in their decision making process and looking for opportunities and tools to fill them. This is an important first step in understanding how to leverage both data analytics and human expertise to help guide decisions in a balanced way. Potential gaps to look out for could include:

• Deciding what and how much data to include.

• Which constraints to apply.

• How to model data-based scenarios.

• How to effectively monitor results.

Once an insurer has identified the gaps, it can begin identifying opportunities to close them. This includes experimenting with different technologies, platforms, data sources and human perspectives to identify the ideal mix. Building in additional time to examine different scenarios and their potential consequences is another important step to refining assumptions and understanding the impacts data can have.

While there is a bit of work upfront, these steps can ultimately help insurers transition toward a decision science approach, which allows insurers to leverage human expertise and data to treat decision making like a science. Doing this not only gives insurers a better understanding of how different decisions perform in the market but helps them tackle persistent and complex problems without the dreaded black box that many traditional approaches involve.

It's the benefits of an automated process without losing the crucial human element. Instead of making the choice for decision makers, a decision science approach to decision making can help ensure that the ultimate decision isn't just a good one — it's the right one.