The State of BPMS Intelligence

20th of December 2017

What is an intelligent BPM system? The term was coined by Gartner as an attempt the categorize the second generation of BPM systems that had capabilities beyond just process automation. These capabilities should give actionable real-time insight into the operational aspects of process executions that can be used in orchestration and adaptation of process instances. In other words, the data gathered from process executions should optimize the new instances being launched by the system, and even transform these instances resulting in better outcomes.

Gartner defines the iBPMS (intelligent business process management suite) as an integrated set of technologies that coordinate people, machines, and things. As with all encompassing definitions, this always struck me as a bit too vague to be of any actual use. But they do go a bit more into detail in their Magic Quadrant for iBPMS published in October 2017. Here they state that such suites should have the capabilities for support of real—time human collaboration, and for advanced analytics, real-time activity monitoring and decision management for intelligent coordination/management of the interactions of process participants. So, in essence, the systems should allow for a management console that allows for the command & control of process instances, combined with a Business Activity Monitor and a Business Rules Management component that allows these data sets to be reused in subsequent process instances.

Gartner does emphasize the solutions as a high-productivity application development tool, allowing for the collaboration of what they coin “citizen developers”, a grouping of business analysts, subject matter experts, data analysts and IT developers, so they do shy away from the zero-coding promise of some suites with this statement. They also state that an iBPMS should have all (or at least as much as possible) the intelligence-related capabilities listed in their Business Process IQ Framework. Basic capabilities in this framework are Data Timeliness, Context Granularity, Predictive Power, Actionability. More advanced capabilities are Contextual Extensibility, Decision Diversity, Optimization, and Behavioral Impact Capture. This framework is then added into their 6 steps for selecting the proper iBPMS for your organization, as shown in the image below.

6 Steps for Selecting an iBPMS (Gartner)

These capabilities are almost all of a quantitative nature, and thus a scale should be decided upon when using them to grade possible products. This is why I prefer the critical capabilities Gartner specifies in the notes of the magic quadrant paper. These capabilities are more like a checklist, and thus easier to decide upon. The solution either has it or it doesn’t. These capabilities are:

  • Interaction Management: The basic workings of a BPMS. Automation/Orchestration of the tasks that make up a process.
  • High-Productivity Application Authoring: Possibility of adapting the process templates that will be used to instantiate an automated process.
  • Operational Monitoring and Business Alignment: Availability of metrics and a tool for the monitoring of business activities (such as for example Business Activity Monitoring (BAM) or Complex Event Processing (CEP)).
  • Business Rules and Decision Management: Integration of a Rules Engine.
  • Analytics: Services that allow for predictive analytics and analytical insights.
  • Interoperability: Capability to interact with other packages over industry standard protocols (REST, SOAP, JDBC, JMS …).
  • Intelligent Mobility: The ability to provide user experience towards multiple mobile devices.
  • Process Discovery and Optimization: Metrics and Analytical capabilities to optimize existing processes, and discover new ones (for example based on desire lines in process execution).
  • Context and Behavior History: The iBPMS should have a log linking events that occurred during process execution and how they relate to the tasks in that process, so that the context of each instance can be reproduced/deduced.

Looking back at Nathaniel Palmer’s take on the modern BPMS during BPMNext this year, the Gartner capabilities seem to be a bit lacking, as they are all focused on the rightmost two domains specified by Nathaniel, namely Process Management and Decision Management. Gartner does not seem to include any RPA/CPA or Machine Learning into the mix of needed capabilities to judge current iBPMS implementations on. For me, these missing domains are the key that BPMS solutions should invest in to be ready for the technological shifts of the near future. Even in his book “iBPMS - Intelligent BPM Systems”, which he published in August 2013 right after Gartner introduced the term, he already expanded on these capabilities.

Nathaniel Palmer's View on Modern BPMS Framework
Nathaniel Palmer's View on Modern BPMS Framework

For the analytics part, Gartner situated the actionability of process data on two levels. The macrolevel where a combination of CEP, BAM and stream data monitoring continually evaluates the process data in order to make intelligent optimizations (both at the real time and the design level). Outlier detection and exception recovery come to mind for this. And the microlevel is more the decision automation and decision support within the process instances that could/should incorporate real-time analytics (such as the goal-oriented BPMNs of the Whitestein BPMS). This is also reminiscent of what Bill Schmarzo (CTO of Dell EMC Service), who came to a similar conclusion about IoT BigData being used for decision making. His approach to turning actionable (big) data into key decisions through Analytics-as-a-Service and providing these insights to intelligent applications, can serve as a template for the analytics capability of iBPMS solutions, with the process instances being the intelligent applications.

Intelligent Applications Stack (Bill Schmarzo)
Intelligent Applications Stack (Bill Schmarzo)

In conclusion, for the near future Gartner needs to make quick work of the capabilities it attributes to their term “iBPMS”, with the inclusion of RPA/CPA (as discussed in a previous thought), as well as how it will incorporate Machine Learning, or it will be a concept left behind in the dust of new concepts racing forwards. In any case, it feels like the prefix “intelligent” will fade, as we will quickly become accustomed to all applications being intelligent, no longer needing to use this prefix for the purposes of differentiation. It will go the same way such prefixes as “electro-“ and “cyber-“ have gone before. Here is waiting for the next prefix with the necessary geeky glee!

Thought BPM