Workflows everywhere: In trouble ticketing platforms, CRM, Inventory Management Systems, Service Management Systems, Order Management Systems. Each of these systems implement some kind of workflow engine to automate the processes they depend on.
Would not it be wise to merge all those in a common workflow engine? Would not it be more easy to manage the workflow environment and migrate to a BPM one?
It will definetly increase managebility and reduce costs. But the biggest obstacle that rises is the integration tax. Centralization means complete integration. Each and every platform should be integrated to this centralized workflow management system via some kind of distributed computing technology such as web service or CORBA. The systems that implement older technologies will not support the newer integration enablers. When this happens, it will be a necessity to install an ESB in the middle.
The second concern that may appear at the application side is the flexibility of those applications. The applications should allow to hand over control logic to an external application, and some of those, especially the legacy ones will not allow you to touch their logic. (The same problem will also be faced when dealing with BRMS platforms) . To overcome this, we may implement “dummy” workflows that will trigger the main workflow engine, but again, you will have to design the workflow in two places.
Because of these reasons, the centralized workflow environment may bring additional burden instead of removing some. For the greenfield type operators , that can be a highly applicable strategy but for operators who has lots of diversed systems and vendors, that seems not be very feasible.