When business processes get large, some additional level of abstraction is required in order to deal with the complexity. One of our customers has a process made of 250,000 activities. Yes, you read correctly, that’s a quarter million boxes connected with each other through very many arrows. When building the system that is automating such a process, we had to put in place some specific best practices supported by a strict methodology in order to cope with the sheer size of the effort. A couple of technologies can help reduce the complexity, by providing some additional level of abstraction that is orthogonal to the flow of the process. Among them, business rules and metadata will play a key role and significantly impact the architecture of next-generation BPM systems.

One vendor we are working with for business rules is Corticon, which Gartner recently positioned into the leaders’ quadrant of their 2005 business rule engine magic quadrant. The reason why I like Corticon’s technology is that it’s one of the few business rule engines that was specifically built for BPM, for a couple of reasons: First, the spreadsheet-like tool is actually usable by business analysts without having to learn any proprietary programming language. Second, the engine is optimized to deal with the large datasets that are usually handled by a transactional BPMS. Earlier today, I had lunch with Mark Allen, Corticon’s CEO, and we discussed about the technical reasons that make their product so much better suited for automated decision processing than any traditional rule engine. If you want to learn more on the subject, read this article.

On the metadata side, I have been following Unicorn for the past four years. This Israeli company started as an ontology company, then moved to the metadata repository space, competing with ASG-Rochade and MetaMatrix, even though the later tends to be more focused on the Enterprise Information Integration (EII) side of the story. What makes Unicorn different is that it blends ontology modeling, schema mapping and metadata repository into a single product offering, doing for data what a BPMS is doing for processes. I had breakfast with Zvi Schreiber, Unicorn’s CEO, earlier this morning and learned a lot more about his product and how customers are putting it to use. This made me wonder how it could be integrated with our BPMS and to what extent it would have simplified our work with the above-mentioned customer.