Part 2: Decision Management and Case Management
In Part 1 of this series, we reviewed how decision management, decision analytics and case management combine in systems that support automated decisions.
Aspects of Case Management: Operational and Investigative
Case management has two key aspects to it:
- Working on cases the automated system could not process or reportedly did not process well (operational or intervention case management)
- Investigate sets of such cases in order to identify how the automated decision could be changed to do a better job and reduce the load on the more expensive case management (investigation case management)
While the distinction may appear arbitrary and there are situations in which the case worker will straddle across both worlds, we can still separate the two aspects for the purpose of this discussion.
In this part, we’ll focus on the first of these aspects. Typically, case workers will be involved for various reasons – for example
- Handling cases where the data is incomplete
- Handling cases which present too much risk for the automated system to take the final decision, but present enough potential for them to be considered by humans, as is the case in a number of loan origination situations
- Handling cases which require interactions with human customers to get to a conclusion on the decision, as is for example the case in credit card transaction fraud in which a call may settle the decision on whether to authorize or not a transaction
Responsibilities of Operational Case Workers
Operational or intervention case workers tend to work with a case at a time, and to review the specifics of the case using forms. They review the case in the representation, and based on their expertise and their hunches will:
- Try to understand why the automated system did not take a decision or why the decision was reportedly incorrect
- Decide what the next step in the processing of the case should be (ask for more information, take a decision for the case, etc)
Decision Management to the Rescue of Case Workers
Decision analytics and modern decision management user interfaces can help.
SMARTS offers the RedPen™ approach to managing decisions. This approach presents the decision information in the context of cases and groups of cases, and lets the case worker interact with decisions within the forms that they are used to in everyday work.
What’s more, SMARTS allows case workers, in order to get support and to track the reasons for which decisions are overridden and/or modified, to use the same system as the business users managing automated decisions.
In the following SMARTS screen in RedPen mode, the case worker looking at the current case (1/100 in Census 2000) immediately understands the reasons for the automated decision (“High Risk” rule fired because “AverageAge” is between 16 and 25 and set the “Assigned Risk Level” to “High Risk”).
Furthermore, the case worker can use the same tool in order to explore how to propose a modification to the decision logic in order to take care of the case being worked on. For example, the case worker can decide that even though the average age is quite young, the total work/school mileage is below 50 miles and that the organization should grant a lower risk level. The case worker can do that by directly manipulating the decision within the tool, creating an exception rule to the “High Risk” rule and changing the conclusions for the cases where, in addition to AverageAge being in the specified range, the corresponding mileage is below 50 miles.
The exception rule thus created may then be fed back to the automated decision, shared with other case workers who can continue refining it. And they can do that with full understanding of the consequences on the various groups of cases and business performance objectives. This is a great example of “design by doing” and adaptive case management.
This approach allows case workers to benefit from the systematic capture of decisions within the system, yet gives them the flexibility to adapt the logic to the situations they are dealing with. It also allows the organization to capture in a systematic way this highly valuable information which otherwise gets lost or is not efficiently communicated.
It is not even required that the case worker actually configure the logic – just highlighting the reasons why the decision needs to be adjusted for this case is valuable information that can be tracked in the decision management tool in order to further improve the decision.