There is sustained interest in underwriting engines, and rightly so because they are in many ways the future of underwriting. Not entirely, for there are and will continue to be risks and situations that engines cannot handle, but automation is the way the game is headed, not just in the UK, Australia, New Zealand, South Africa and increasingly North America but in parts of South-East Asia too. Continental Europe remains a bit of a curiosity in that deployment of engines has been a bit patchy, certainly for life insurance, although much health insurance business is auto-underwritten in Germany. Even in parts of the world where risk assessment is still extensively manual they are looking at developments in automation elsewhere.
If you are moving towards automated underwriting, what tool or tools should you be looking at? There is perhaps a bewildering array of choice in terms of products and related components. What you need depends to a large degree on what you want to do. Some form of simplified issue, or something more sophisticated? Which channels will the new processes be applied to? What components in the new business pipeline do you have in place already?
Do you need simply an engine as a ‘black box’ decision maker, or do you need additional software to form the front end/user interface, design forms, manage the process, etc? The answers to these questions will determine which suppliers you need to be considering. An important question concerning suppliers is whether you want your engine to independent of reinsurance. Each of the big international reinsurers offer an engine and most of their products are good. Most likely they will sweeten the engine deal if it is linked to a reinsurance treaty, but how much do you value your independence? And given that certain reassurers have a thirst for data these days, how much are you willing to share the data generated by the engine if that is a condition of the arrangement?
Then there is the matter of where the software sits. On your servers or the supplier’s? There are proponents for both approaches. Those in the ‘we host’ camp cite control and security of customers’ data. The ‘hosted’ fans would say it’s simpler and quicker (and we confess that we do like the idea of making the IT department redundant apart from looking after the hardware and providing user support).
It is worth considering carefully the functionality on offer from the various suppliers. There are plenty of decision-making engines out there, many at what appear to be very reasonable prices, but how many have been developed with the life insurance industry in mind? ‘Universally applicable’ engines have to be developed from scratch, which means a great deal of time and effort and the risk of an inferior set-up at the end of the project. They might look attractive because they are low-cost and can be/are being deployed elsewhere in the business but beauty is only skin-deep, as the old saying goes. Underwriting managers should resist the proposals of the IT department (who’s running the business anyway?) and hold out for a purpose-designed product. In the long run that will mean a better, swifter result at lower risk.
But even purpose-designed products vary a great deal in what they offer and what they do. For example, some come complete with a set of rules that can be customized, others don’t. The ‘with’ products are not necessarily inferior to the ‘withouts’: views vary on whether it is better modify a basic rule-set or to start from scratch. Much depends on the quality of the rules a product comes shipped with and the environment in which the engine will operate. Sometimes it does make sense to start from base one.
Functionality is another key issue. Fact is, some engines are easier to use and manage than others. We are firm believers in IT working for the business and that users should be able to implement and maintain an engine without programming skills or the need to involve any pointy-heads. Using it should be as easy as working with anything in the Microsoft Office suite, but there are products available that are not user-friendly, need a significant amount of training and have a distinctly 1990s look – clearly they have been starved of investment and are best avoided.
Modern engines have one straightforward-to-use dashboard from which all the various elements can be managed; for example:
- Decision trees visualized and rules created and modified easily by dragging, dropping, copying, cutting and pasting
- The ability to add features that contribute to the customer journey such as help text and header/footer text
- Flexibility as to how dates are entered – otherwise rules developers are restricted in the format of questions
- ‘Spellchecking’ or recognition of mis-typing, particularly of medical condition names
- Strong version control
- Integrated testing functionality.
Given that a key benefit of an underwriting engine is the generation of management information, check what comes as standard. Does the management information module come at extra cost? Are there any standard outputs? How easy is to create additional ones? Does analysis and interpretation of data require buying additional software? Can data be easily exported to other statistical software programs your company already uses?
It’s not exactly a minefield but you need to choose with care. The wrong decision can make life more difficult than it need be.
