Adaptive Dynamic Processes or just plain work…

Gartner’s Janelle Hill recently commented that “By 2013, dynamic BPM will be an imperative for companies seeking process efficiencies in increasingly chaotic environments”. Max Pucher proposes a definition for Adaptive Processes and Peter Fingar (who I regard as one of the “real” thought leaders) support Keith Harrison-Broninski’s Human Interaction Management concept in his Work 2.0 article (highly recommended reading). Keith Swenson, VP of R&D at Fujitsu America Inc., did an excellent interview with Jean-Jacques Dubray whose publications I’ve been following from my early BPM days. In the interview Keith gets Jean-Jacques view on “Does unpredictable work exist?” and whether you agree with Jean-Jacques that no process is unpredictable, there is certainly a renewed interest in the seemingly chaotic way we work and how to manage it better.

One of the comments on the interview states that “… to start a journey, a path must be chosen. All you need is a map, maybe even a compass will do”. It continues that the point is that the end path is not always known ahead of time. This is certainly true for a large component of the work that we (knowledge workers) do on a day to day basis. Our work has changed from the factory style, sequential and pre-determined workflow, to dynamic event based process “reactions” based on the contextual knowledge that we have for the activity that we need to address. This is quite a long explanation. Let me simplify with an example.

Suppose we receive an email from a client that want to change a purchase order from a 1 000 widgets to 10 000 widgets. Their initial order was wrong as someone left the last digit off. They still want it on the same delivery date. Suddenly there are a number of things that need to happen in your process. Our hard-coded predefined process requires that the sales person completes an “order variation” request that will be loaded in the ERP system. It may now move the delivery date out due the large variation because the MRP planning module and it may also moved to pending status as the customer credit limit is exceeded and needs to the adjusted. This is where the sales person generally revert back to email and collaboration as the structured dependency of the process limits his or her ability to “manage the process effectively”. The credit change process is manually escalated early on while the Operations Manager is called/emailed and briefed on the possible change and impact on inventory and production planning. Existing process controls are bypassed and process actions are dynamically routed and escalated based on the information that we gather at each step. We may find that credit approval requires additional information and possible collateral from the customer and we “know that we are about to receive a number of additional orders from the client as our account manager informed us that the customer landed a big contract”. Do you see the complexity in the process? It is not in the actual execution of each task, it is in deciding on a course of action based on information that we gather in the context of the process or as it emerges. It is not the activities but the combination and permutations of when and how we need to execute them. The above example is a simple one but one that most organisations would address collaboratively if they had to do it right now.

We generally model the suggested or “happy path” but this is not to say that it is the only effective way of doing work. Sometimes (most of the time) we need controlled process activities that we can sequence as the process requirements “emerges”. That is an Adaptive Process. It adapts itself in each specific transactional instance. We still do the same work, but we done necessarily do it in the same way every time. We still have specific process activities that are compulsory in every process, like credit approval, but we sequence it based the dynamics of the business. This is true “process management”.

The above example demonstrates the challenges that we face with models that are based on sequential workflow style routing and I’ll be posting on follow up posts on some ideas and suggestions around dynamic models.