Jim SinurI may be dating myself, but the lyric “Tommy can you hear me?” comes to mind when I think of listening. This lyric came from the rock opera Tommy by “The Who” where a pin ball wizard was deaf and blind. It is a good listen with a bottle of wine, but that is for a different time and place.
Processes have to learn to listen better moving forward, so I would like to outline some of the ways BPM can do a better job. Processes must do better than depend on process participants and managers to catch when critical, important and interesting events are occurring during the execution of process instances across a process. Processes should notify and potentially fix issues automatically.
![MP900382872[1]](http://blogs.gartner.com/jim_sinur/files/2011/08/WindowsLiveWriterNextGenBPMEncouragingYourProcessestoList_89B9MP9003828721_2.jpg)
Listening Should Occur Throughout the Process Scope:
This means that processes should be listening to the end to end process for possible issues with the expected results of a process. If there is a set of expected results from the process, the process should listen for the tell tale events that could risk those outcomes. the simple approach would be to listen for out of tolerance conditions. These conditions could include measures that sound the alarm or constraints that should not be violated by a dynamic process. Processes should listen for resources that are becoming constrained over time and predict potential poor results ahead of time.
Listening to the Context of the Process:
It’s not just the results of an activity within a process than can sound off for attention, but it is activity around the process. A process can listen to the behavior of a market and let those in a position of authority that there is market behavior that needs to be acted upon or further watched. The same can be said for other contexts such as dynamic value/supply chains, geographies, government rules with expected changes and other inter-organizational dynamics.
Listening for Expected Events:
This can be done by listening for simple single events and/or complex events. This is where there may be a set of predetermined rules looking for event pairs and/or combinations that are of interest to process participants, managers and/or owners. This makes processes listen in “real time” and generally requires a significant complex events processing (CEP) capability.
Listening for Unexpected Events:
Quite often business exceptions end up washing up on the shores of business processes. Looking at where these instances get collected is how these are reactively found through some form of reporting. Proactively organizations can set off alarm bells by adding some smart tolerances that can catch when things go beyond normal expectations.
Listening to Suggestions:
Processes should be instrumented with the ability to collect suggestions from process participants whether they be employees, clients, vendors, and other various constituents of a process. Smart organizations also look to the behaviors of the collective potential clients etc. to get ahead of the curve.
Just like in life, good listening skills go a long way to successful outcomes. Processes have to mirror this behavior to stay sharp and on top of outcomes before they arrive on your business shores.
Category: Applications BPM Business Process Improvement Business Rules Cloud EA Optimization Simulation Social Strategic Planning Virtualization Tags: BPM, Business Rules, Decision Management, events, Optimization, Process Improvement, Simulation, Social
4 responses so far ↓
1 Philip Hogg August 16, 2011 at 6:51 am
Your thoughts in the area are very timely for me. We are looking for ways to get the earliest possible alerts that things may not be working as they should for vulnerable families and particularly children at risk. We want to shift our service focus from reactive to preventative.
As you suggest we need to look for the desire outcome and monitor for deviations, but not only in our processes but also in the processes provided by others.
In our complex human services area it is not easy to identify the deviations. Some deviations may be small but the accummulation of small events across a number of areas may provide an early warning.
Has anyone else done some work on early warning systems in the humans services area?
2 BPM Quotes of the week « Adam Deane August 20, 2011 at 3:21 am
[...] BPM and AI – Jim Sinur Processes must do better than depend on process participants and managers to [...]
3 BPM and Artificial Intelligence « Adam Deane August 22, 2011 at 3:14 am
[...] was reading a blog by Gartner’s Jim Sinur on the topic of Next Generation BPM. Processes must do better than depend on process participants and managers to catch when critical, [...]
4 Max J. Pucher - Chief Architect ISIS Papyrus Software August 22, 2011 at 5:20 am
Jim, I am glad to read that we were quite aligned in our discussions recently about the future of process management. I find your description of a ‘listening process engine’ intriguing. As I mentioned in our conversation I propose to expand ‘listening’ with ‘learning’, meaning that the process engine is be able to learn by linking events and user actions to process patterns as does the pattern matching machine learning component of the Papyrus Platform.
Your ‘listeners’ we perform as follows (they are all really the same):
> Goal-oriented processes and boundary rules (constraints) automatically add activities on mismatches, triggering corrective actions.
> Business and market data can be federated into the context of a case in real-time, which the user-trained pattern matching and/or business rules use to identify trigger conditions for action.
> Expected or unexpected events are never standalone but it is the context that makes them relevant or not. Therefore the engine looks for event/context patterns and not just the event trigger.
> When actors perform any kind of observable action within a case/process (which is in fact an EVENT too) then the pattern matching agent will create an event/action pattern and if repeated recommend that action to other user roles in similar situations. Rejecting or accepting provides further training to improve the engines suggestions. Users can also decide to allow the agent to automatically execute certain actions above a probability threshold.
This pattern matching agent was patented in 2006:
http://www.faqs.org/patents/app/20110178965
Thanks for finally bringing the future of BPM to the attention of the community. Regards, Max