Business Intelligence and Business Processes - James Taylor's Decision Management

« More on decision services | Main | Great article on Agile Compliance »

March 20, 2007
Business Intelligence and Business Processes

I saw a Doug Henschen post recently - Make the BI/Business Process Connection - and it prompted me to reiterate some of my thoughts about the power of decisions to integrate BI and BPM. Doug had been talking to Gareth Herschel of Gartner. I have blogged before about Gareth's excellent presentations on this topic (in particular here after attending an excellent ITxpo session of this). Doug talks about an example Gareth shared - ING Insurance who used "historical and predictive insight to screen claims, fast-tracking the many claims falling within expected parameters while giving the few exception items greater scrutiny". This is a great example not just of brining intelligence to business processes but of embedding decisioning in a business process. ING automated the decision as to whether a claim should be paid or scrutinized. I am willing to bet this took rules (regulations, policies) and analytics (predicting the risk of fraud for instance) or what some call EDM (Enterprise Decision Management). What ING did was make their claims system smart enough to recognize most, though I suspect not all, of the valid claims and process them straight through.

Doug quotes Gareth as listing ten common BI and analytic techniques: "simulation, optimization, events alerts, reporting, dashboards, visualization, ad-hoc query, statistics, descriptive data mining and predictive data mining". Hmm. Me, I like the old Thomas Henry Huxley quote:

"The great end of life is not knowledge, but action"

So I am less keen on reporting, dashbaords, visualization, ad-hoc query, statistics and descriptive data mining when it comes to improving business processes, and more keen on the active ones:

Predictive data mining (or predictive analytics) to turn your data into actionable insight as you need to use your data not just collecting it.
Simulation to see how different models (and rules) would make for different decision outcomes
Optimization to come up with the best possible rules for a given decision
Event alerts, but only if you can do something not just display some data (see this post on decisioning and BAM)

I believe that you need to think about a decision management platform because only decision technologies can be a platform for analytics in BPM. This allows what some at Gartner call a transaction-centric process and enables you to put your performance management into action.

Sorry I have been lax posting, had a deadline...

Technorati Tags: BAM, BI, BPMS, BRMS, business process management, business rules, data mining, decision service, EDM, Enterprise Decision Management, Intelligent Enterprise, performance management, predictive analytics