That said, I do think people always need to be in the loop. The computer only does what you tell it: given a goal and some information, it will make choices which maximize the goal. But somebody needs to make sure the computer has the right things to choose from, is getting the right information, and has the right goal. Indeed, these things change over time (often fairly rapidly) and people will need to monitor any decision system to ensure it’s behaving in the intended manner.

Dave

In other words, an intelligent business is a very different beast to a business that uses business intelligence.

Cheers,

Rob.

You’ve confused me for Curt Monash. I run the Monash BI Blog which is a blog of the BI research group at Monash University (in Australia). Confusingly, Curt is well known in BI circles, so the mistake was bound to happen at some stage!

I’m the author of the Myth of BI for the Masses post.

A great write-up though. I agree that there is probably some scope for embedding some decision making processes within systems (this happens anyway though, with any systems design - business rules have always been part of systems analysis). I’m somewhat skeptical of the possibility of full automation - and wary of the problems when the system gets it wrong. Claims from the AI community have been along these lines since the 1960s (Herb Simon famously claimed that we’d all be living in a bucolic paradise while machines did all the work by 1970).

No system can make decisions as well as an informed, competent human decision-maker. There’s also ethical issues associated with accountability and moral responsibility for individual decisions if no human was involved.

In any case, I do see a big difference between operational reporting needs (these are much easier to solve) compared with strategic decision-making needs. The BI vendors, generally, though see their tools as a universal solution. It’s a view based more on a desire to sell software licenses than on an accurate understanding of the cognitive and organisational factors involved in making an important decision.

Cheers,

Rob.

Perhaps much of the organizational struggle around how decisions are made grows from the often denied or ignored reality that we don’t agree on what problem or business issue we’re trying to solve by the decision. Until people agree on the problem and its cause, it is unlikely that any amount of data or robust process will result in timely and effective decisions.

Paul Nutt’s research from Ohio State (Why Decisions Fail) presents an overall dismal view of the effectiveness of most corporate decisions.