I’d like to point out that there is a limit to decision services or decision management or decision automation, whichever term you use. When James Taylor and I wrote Smart (Enough) Systems, we were very careful to point out in the book that this applies to only a very limited number of decision types, not all decisions. Those that can be codified are generally customer-facing (though there are others such as logistics, etc.), generally of large numbers and, most importantly, not so critical on an individual level that it hurts too much to be wrong. In other words, if you get 100,000 insurance inquiries a day and deny coverage to a handful that should have been covered (or alternatively, cover some that shouldn’t), it’s OK because you keep working at it to get better models (adaptive control).

Decision services can also advise humans where it would be reckless to allow a decision model to take complete control.

Then, of course, there are decisions that are best left to experts.

Sorry I couldn’t make it to BRF, I was elsewhere.

-Neil Raden