As an ECM solution implementer, the concept of Case Management has become more and more prominent over the past few years. Within the FileNet community, Business Process Framework has been the application of choice for case management since 2005, however this has slowly been erased from IBM’s ECM offering since the turn of this year – you’ll do well to find any sales or product literature for it on the IBM website. The reason? BPF now has a successor: IBM Case Manager. Apparently, BPF will still be supported as a product for quite a few more years, but it’s obvious that as a proposition it’s been pushed to the side to make way for the new kid on the block.
Inspired by BPF, ICM was released last December. IBM have certainly learned from their customer experiences in case management. As well as replacing BPF, it’s also positioned to address the demand for a more natural approach to case management – ACM – Advanced Case Management (a.k.a. Adaptive Case Management / Dynamic Case Management).
So what is ACM? Well, it introduces a spectrum of business processing, in the context of structure, from one extreme of well-defined procedures, to the other extreme of unstructured processing. It’s targetted to knowledge workers, who are distinguished from typical BPM agents by their ability to handle unpredictable work, allowing a case worker to use their own knowledge and experience to determine the next steps to progress the case towards a successful outcome. The crux is that instead instead of each case type being associated with one process, multiple process fragments can be defined, with the knowledge worker being empowered to optionally add any of these ad hoc.
While getting to grips with ICM, it’s impossible to escape the conclusion that it bears some striking similarities to BPF’s case model. But unlike BPF, ICM extends from the core of P8 and is completely integrated, instead of being built on top of it; the case management objects are now native to the Content Engine, as opposed to being simple custom objects. FileNet P8 v4.5.1 introduced the idea of Roles and Inbaskets, and while it seemed this was aping BPF, it’s now apparent that this was merely paving the way for ICM. ICM is well abstracted, I guess to become more appealing to Business analysts, and does a great job of wrapping up the underlying storage within CE and PE.
A rich development environment is also provided – The Case Builder – a tabbed collection of wizards and widgets. All solutions are preserved withing a Design space, and when complete are easily deployed to the online target space, ready for testing in the Case client. A Case Deployment facility is also available to promote from the development environment to test and production.
Tasks are input using the Step Editor, which is based on a swimming lane model. Steps are very basic, however the more technically-orientated can refine these in the Workplace XT Process Designer, since the tasks are stored internally in XPDL format. Obviously, support for Process Designer provides the opportunity to enhance and customise the process definitions through the Java Component integrator interface. User Interface customisation takes place within the Mashup pages through iWidget – this is an open standard, and there have been plenty of tutorials available describing how to create bespoke widgets ever since ECM Widgets was released. The Javascript/AJAX library Dojo is also part of the make-up. For anyone unimpressed with widgets, preferring an alternative web front-end, the case management data model is exposed through three REST service APIs: Case REST, PE REST & CMIS REST.
The big worry for the existing BPF customer base is that there is presently no migration path to ICM. Despite being labelled as the next generation of BPF in some quarters, it’s an entirely different product line.