Are we reaching a point where many of our corporate decisions can be automated? Are we ready to make this leap?
At some level, we already rely on software to make minor, everyday decisions. Heres a classic example: If you use Microsoft Word, the software auto-corrects spelling for commonly misspelled words and capitalization. We leave it to the software to make these low-level adjustments for us, and these adjustments are based on common rules. Of course, Word wont create the document for us, thats still the role of a human decision-maker.
Likewise, with many enterprise systems, such as policy administration, underwriting and call center systems, there are countless low-level decisions that can be effectively automated. I recently had the opportunity to speak with both James Taylor and Neil Raden, the co-authors of Smart (Enough) Systems, about where decisions can be effectively automated.
Taylor, for one, advises companies looking into automated decisioning to automate the really obvious decisions. The evolution to automated decision making, he said, is highly incremental.
Its very unusual for these systems to jump straight to 95% automation, he says. Its much more common for them to start at 50% to 60%, and then review manual decisions each day or each week. Gradually you start to understand that whenever certain things are true, we always do this in response, so we add those rules. So eventually, the percentage rises.
Raden observes that in the insurance and financial services sector, decision processes have been hard-coded into legacy systems that have been built up over the years, and this has created inflexibility.
Banks, credit card companies and insurance companies have been building what you could call decision management systems for years, regardless of whether or not they used rules engines, he says. But they likely just hard-coded the rules themselves. Thats a real killer in terms of productivity.
A separate business rules engine could abstract business rules from the decision process, and can be adapted as the business changes, Raden says. Those rules could be managed by people, not IT. They could be used over and over again in a whole bunch of other things. Consistency in decision-making is the advantage to be gained by such systems, he adds.
In Smart (Enough) Systems, Taylor and Raden provide illustrations of how automated decision management can reshape insurance operations. For example, a company may have a 30-year-old claims processing system with hard-to-decipher decision logic. An automated decision management approach may consist of creating decision services at a centralized environment that would handle claims adjudication and payment among other decisions.
The centralized decision services would provide consistent decisions and benefit calculations for all claim types, and enable a high rate of auto-adjudication for maximum efficiency, they write. Decision services can apply additional rules to route claims or the correct departments. These same rules deliver Web-based self-service for claimants.
Underwriting systems also tend to be hampered by manual processes that often result in slow and inconsistent decisions. Underwriter time gets used up, and its difficult to keep track of reports ordered. A centralized decision services can combine business rules with risk models with predictive analytics.
Business rules ensure that external data is brought into the decision-making process only when it makes a difference, they write. This method speeds approvals, increases straight-through processing, and reduces costs ... Quote counts and new business has increased because agents can write policies immediately.
Joe McKendrick is an author, consultant, blogger and frequent INN contributor specializing in information technology. He can be reached at
