5 Business Reasons to Service-Orient and 3 to decision-automate - James Taylor's Decision Management

« Making the case for rules in compliance | Main | IBM, SOA, Second Life and blogging »

December 15, 2006
5 Business Reasons to Service-Orient and 3 to decision-automate

I got a note from the folks at ZapThink about an article in SAP Netweaver magazine called "5 Business Reasons to Service-Orient". The 5 reasons given are

"reduced integration costs, automated business-to-business interaction, easier regulatory compliance, user empowerment, and the ability to react to change more quickly"

The article goes on to use an illustration of the way "Carla" works to show this. In fact Carla would only really become more effective if the company not only adopted a service-oriented approach but also automnated some key decisions. Let's consider the scenario:

"Carla is a call-center representative for a bank. She takes calls from banking customers and accesses various systems on her computer to help address each customer’s question or concern"
Clearly Carla is being asked to make decisions about customer treatment all day

"Despite her best efforts, Carla occasionally makes mistakes that she or another call-center representative have to fix later"
Some of these are probably data entry errors and so on but others will be decision errors - refunding something she should not, allowing an account that is not allowed, failing to flag a suspicious transaction, using the wrong pricing model etc.

"Although her boss has repeatedly told her that an important part of her job is advising customers about products they might be interested in, she has a difficult time understanding which products are right for each customer."
Hardly surprising given the number of products offered by a typical bank (which have all the signs of a classic "Long Tail" business)

"After her bank implements SOA, however, Carla’s daily life becomes much simpler...Now, she can spend less time on each call, and information about her customers is at her fingertips. She can identify at a glance which of the bank’s products will be of most interest to the customer, leading to increased sales for the bank and a bigger bonus for her. "
Well perhaps but probably not. Simply having the information at her fingertips won't necessarily help her pick the right product/cross-sell/up-sell from the hundreds available (one major bank recently launched 300 new credit products in a month, for instance). To do this she would need to combine not just the data but predictive insight from the data (this customer is likely to accept this kind of offer) as well as a bunch of regulatory and policy rules in the context of a strategy she cannot know in detail. Ain't gonna happen. The reality is that unless this kind of decision-making is automated in a way that leverages the insight and enforces the rules and gives her recommended actions to take, she it not much better off.

"Furthermore, as various business changes work their way into Carla’s inbox, she can make adjustments quickly and efficiently. Maybe it’s a new marketing campaign followed by a regulatory change, or maybe a new product the bank wants to promote. Service-oriented tools are so powerful, flexible, and easy-to-use that Carla can take advantage of them to do her job regardless of the changes her boss throws at her."
Now why on earth would you want to email Carla new marketing campaign information, new product information or new regulations? Email? She could get dozens of emails a day on these topics and they would likely be buried in hundreds of work ones. There would also be no way for her to tell which ones replaced which other ones and no way for her management to check that they did not contradict each other. And what if the new marketing campaign was something like "offer this new high profit, high risk product only to customers who are not currently profitable and have a risk profile of such and such"? How is she going to do that risk assessment? What if the regulation was BASEL II and meant she had to make decisions based on an updated credit risk model or if the FBI was sending new warnings out about money laundering that she was meant to enforce? Would you really want to tell all these regulators that you were enforcing these regulations appropriately and that you knew you were because "we sent an email to everyone telling them"? Don't think so...

The reality is that for many businesses that are only going to get better regulatory compliance, user empowerment and business agility from SOA if they also think about automating critical decisions in a way that allows users (internal users or external customers) to work without having to constantly refer business decisions to someone and if the way they implement these decisions supports both rapid change (which means business user change not IT change) and regulatory compliance (you can prove to the regulators that you are doing it right). This means business rules as the basis for decision automation and predictive analytics to bring insight to bear on these decisions.

For a couple of other examples of how decision automation can improve a bank, check out this banking story and this one on using decisioning to build the bank of the future.

There's more on regulatory compliance, BASEL II, money laundering, Sarbanes-Oxley and business agility on my other blog also.

Technorati Tags: analytics, BRE, BRMS, business agility, business rules, compliance, customer insight, decision automation, long tail, predictive analytics, Sarbanes-Oxley, SOA, SOX, Netweaver, ZapThink