Final session today is Sandy Kemsley talking about Case Management. Sandy began by considering the system we use to work today. How we work, she says, is based on two different approaches illustrated by Taylor’s theories of standardized processes for efficiency and Drucker’s management by objectives theories. We need to support both and that drives the kinds of systems we need. We need systems for:
- Routine work
Efficiency,accuracy, process improvement and automation.
- Knowledge work
Flexibility, assisting human knowledge workers, collect artifacts.
Adaptive Case Management
Now I would add that different companies consider the same process as either routine work or knowledge work (handling claims, for instance, can be automated routine work OR case-based knowledge work).
Processes can be structured or unstructured or a mix – structured processes with more ad-hoc exceptions or a largely unstructured process with highly repeatable pieces. These mixed processes are much more common than simple structured or unstructured ones in her experience. This variation drives different kinds of technology based on the predictability of the process which varies from very predictable automated processes to human process management to production/adaptive case management to social and completely unstructured collaboration tools like email. All of this creates complexity in the market as different vendors use different terms, package up different capabilities and so on.
Sandy sees three main usage scenarios:
- Process orchestration
Transaction-centric, mostly pre-defined
- Complex knowledge work
Content-centric with pre-defined tasks that the worker picks from
- Simple task/process management
Goal-centric, checklists created and assigned by the worker
Plus she identifies both rules and analytics as adjacent technologies. Predictive analytics as well as mining for process discovery are increasingly powerful analytic approaches while rules have a role to play both in-line to make decisions (Decision Management) and declaratively for triggering actions. In the context of a repeatable process you make an explicit request for a decision and that’s how rules get used. But this is less clear in an unstructured environment – the rules, the decisions, have to be triggered as the case evolves to create guard rails and enforce compliance.
Sandy therefore splits the focus here between BPM and ACM
- BPM, Business Process Management
Repeatable, transaction-focused, efficiency and straight through processing are the objective.
- Adaptive Case Management
Unpredictable, knowledge-centric, problem resolution and support are the objective.
And many processes need both.
When knowledge workers don’t have case management system they have several challenges:
- They sometimes don’t get a system because the perception is too squishy to define.
- They may have overly rigid process that drive lots of manual work-arounds.
- They may lack context for their decision-making, wasting time looking for information in multiple systems.
- Finally they may be inconsistent in their application of policies and rules, with large training costs being incurred.
A robust case management environment then needs to support:
- Content management for the information that makes up the case
- Checklists for the worker to let them track progress
- Processes or process fragments that can be applied to the case
- Events to inform the case and case worker that something outside has happened
- Rules to constrain and fence the behavior of someone working on the case to keep it on track
This means a platform for case management needs:
- Process, content, rules, event, collaboration, analytics
- Persistent case folder
- Pre-defined checklists
- Guidance and guardrails from rules
- Predictions and what-if scenarios
Sandy walked through some use cases to show why its important that a case management capability is integrated with structured process management so that one environment can handle the common scenarios that use both. For instance handling problems with documentation in an automated loan process requires case management and should remain inside the system so that compliance can be demonstrated and patterns be spotted.
Benefits, she says, are improved visibility, increased productivity from better informational context, more practical task reassignment and improved audit and compliance. In addition the use of this technology might allow real-time event monitoring, integration of predictive analytics and mining for emergent processes.
Although Sandy differentiated between decisions in BPM and rules as guardrails in Case Management I think that Case Management needs decisions as guardrails – is this case ready to move to a new state, is this an allowed action, can this task be assigned to this person – these are DECISIONS not rules. They are decisions triggered by events or state changes not by a process step but they are decisions nonetheless.