Step 9 of 10 Steps towards the Agile Enterprise

Till this point in our story, we have seen how to manage Decision automation, policy management, and transfer of Business Decision control & authority back to the business.

A Business is a business after all. The question that will be asked next is this !

We have our business goals to meet. We try and formulate policies so that we are able to meet these business goals.

Now, we have an automated Business Decision Manager.How can we analyze, tweak, improve our policies so that we can pro-actively optimize our business policies to enable us to meet our stated goals?

An example will help understand the point here.

Consider a Mortgage Loan scenario. Now, one stated goal will be to minimize risk for each approved loan. Another goal will be to maximize the amount of loan disbursed.

So effectively, the goal reads as “Maximize the loan amount disbursed while trying to minimize the average risk factor“.

Great! But how will you help your business team formulate policies that will help them in meeting these goals?

Sure they know the business, but we need to provide them with the correct tools and technologies which they can use to solve these problems.

What we need here is the Capability to analyze rules for business target/goal and suggest the best policy formulation that is possible.

There are some other questions that I have heard that beg the availability of a rule analysis capability. Here is one of those.

Now, you need to make a policy change. So, how would you measure the business impact of this change? Traditional regression tests can only indicate whether this change will not break current use cases.

I need to insist here that we are not talking BI, but Policy Intelligence. I use the term Rule Analytics to indicate Policy intelligence here, because this is how we think of it in our company. Now, lets see what it should take to analyze rules for business goal match.

1. A business analyst must be able to work with the business rules by himself without IT’s help
2. They need to be able to take existing policies and create analysis scenarios for the policies
3. They need to be able to simulate these various analysis scenarios without the help of IT, and may be even without an underlying programming model.
4. They must be able to specify a business goal during the analysis/simulation, then specify decision parameters, and business metrics and use the simulations to understand, improve and optimize their policies
5. The business data for the scenarios, they must be able to use the available historical data from the warehouse(wherever it is legal to do so), or from Excel spreadsheets
6. They MUST be able to capture new policies without an underlying programming model simply by using the appropriate business terms, vocabularies & data dictionaries

I suppose these are some broad requirements/capabilities required to undertake a Policy/Rule analysis and gain Policy Intelligence.

Now, to finally answer some questions about Rule Analytics. These are questions that I had when I was introduced to the subject.

This sounds terribly like plain Rule testing? So what is the big deal?

Tests only give you Yes or No results. Rule Analytics gives you suggestions. If you have 30 analysis scenarios, and decision parameters and metrics, Rule Analytics can give you a ranked result of the most optimum combination of these values to be used in your policies to meet your business goals

Now, How different is this from What If Analysis?

A What If is a manual, painful, and therefore error prone procedure to undertake to arrive at the results that Rule Analysis can give you in an instant. Typically, with a what if, you get best case, worst case, and the optimum case is typically guesswork. Rule Analysis takes the guess work out of the equation and helps arrive at accurate precise Business Policies.

Finally

This is it! We have come a long way from basic process automation, to decision automation, to transferring business policy control over to the business and finally now to Business Rule Analytics.

There is one final step left. The final step is more of a future gazing. You watch this space.