James Taylor's Decision Management

Reading Sandy's blog, as I always do, I saw her mention of this Gartner podcast. This made me think about decision management in insurance so I though I would write some more about it.

Insurance has seen a huge growth recently in the move to adopt a decision automation approach. the reason for this is quite simple - insurance people think in decision-centric terms. They recognize that the underwriting decision, pricing decision, claim payment decision and so on are key to profitability. In addition they are trying to increase the number of transactions that can be managed "straight through" and this means automating these decisions. To make these decisions well, and automate them in the process, insurers are adopting two key technologies in combination (as noted in the podcast) - business rules and predictive analytics. The combination of these is known (to some at least) as Enterprise Decision Management.

The technology helps insurers solve a number of problems:

• Collecting the right data the first time is hard because different data is required depending on the circumstances and so as data is entered it changes the remaining data required. It is not unusual to go back and forth multiple times to capture the required data - costly and time consuming. Rules-driven user interfaces, taking advantage of smart forms and AJAX combined with automation of the decision help ensure this is "once and done".
• The decisions are heavily regulated and so not only must each decision follow a large number of guidelines and policies, it must also be possible to show how the decisions taken conformed if/when a regulator asks. To make things more complicated, these regulations are often different by state and country. Business rules technology allows this to be managed and then reported on.
• Tiering and segmentation are key to pricing and risk management. As this has become more and more sophisticated and as more and more data is available it has become impossible for an individual to consider all aspects in a way that allows fine-grained segmentation. Predictive analytic models can automate this complex risk assessment and allow for much finer-grained decisions.

These problems are often found in less decision-centric industries and these are adopting the technology more slowly. It's a pity as the insurance industry, not typically an early adopter, is really showing the benefits of automating these key decisions.

There is more on automating decisions in insurance on my other blog where I also posted on the value and usage of location intelligence.