Today I was able to get hands on with Pegasystems Decision Strategy and Next Best Action Advisor capabilities as integrated from their Chordiant acquisition and I have to admit the offering looks pretty solid. The capabilities are integrated into workflow and are able to access rules readily in the core environment providing for significant leverage of the baseline capabilities and rules inherent in the PRPC platform. A number of new rules have been created to wrap the decisioning capabilities and ease the integration into PRPC applications in such a seamless way as to convince the developer that these capabilities are native to the platform.
The basic premise of these new capabilities could have been conveyed more clearly in the sessions, however it seems fair for the presenters to have assumed people had done their homework up front – sadly I had not so the session was an interesting journey. In the end, my simple understanding is that the decision strategy capabilities are backed by a fairly strong analytics engine to draw correlations between points of data and decisions taken by involved parties to rate the viability of any given course of action in real time based on the party’s data profile and the established correlation model. The correlations are recalculated on a periodic basis as configured by the developer while the data elements are also fully configurable based on the analysts assessment regarding which data may be relevant to arriving at a decision. The overall approach limits the level of knowledge required by operational users regarding the analysis involved and makes much of the configuration manageable by business users with a moderate level of technology sophistication.
The training sessions were pretty straightforward with the general configuration mostly entailing flow-like edits and data table configurations departing somewhat from the more sophisticated and arguably more complex approach of exposing core functionality as business rules taken by Pegasystems in building most of their framework offerings. Considering the core capability is an analytics engine that is likely beyond the grasp of most developers on the platform, this seems to be a very reasonable and well advised approach.
I was somewhat disappointed to see the pure focus on sales and marketing during the training. This focus appeared to go beyond training and into the object models themselves – although a liberal interpretation of the naming conventions could allow for more widespread leverage beyond sales and marketing functions. There is definitely some significant potential here to leverage this technology to advise on next best actions that could save hundreds of thousands if not several million dollars in a single day for certain exceptional operational processes. More abstraction of the core capabilities with a sales and marketing overlay would make the application of this capability to other horizontal functions less of a stretch – but I’m sure we’re going to see this refined in future releases.
There is still some work to be done with the PRPC integration of the decision strategy functionality – the smart prompting for values could be a little more complete, however this is not a show stopper as the expression builder is used in most cases where I wanted the smart prompts and was perfectly adequate to get the job done. The one other noteworthy shortcoming appeared to be a gap in the use of Message and Field Value rules to ease or enable more rapid localization of solutions. This is only a factor in cases where an organization is going to jump in with an international launch on the decision management capabilities – and even in such a low probability scenario there are ways to work around the gaps, although they do introduce certain additional maintenance overhead.