The demo program kicked off in the afternoon, with time for three of them sandwiched between two afternoon keynotes. Demos are strictly limited to 30 minutes, with a 5-minute, 20-slide, auto-advancing Ignite-style presentation (which I am credited with suggesting after some of last year’s slideware dragged on), followed by a 15-minute demo and 10 minutes for Q&A and changeover to the next speaker.
SAP: BPM and the Internet of Everything
W4: Events, IOT, and Intelligent Business Operations
Continuing on the IoT theme, Francois Bonnet talked about making business operations more intelligent by binding physical device events together with people and business events in a BPMS. His example was for fall management — usually for the elderly — where a device event triggers a business process in a call center; the device events can be integrated into BPMN models using standard event constructs. He demonstrated with a sensor made from a Raspberry Pi tied to positional sensors that detect orientation; by tipping over the sensor, a process instance was created that triggered a call to the subscriber, using GPS data to indicate the location on a map. If the call operator indicated that the subscriber did not answer, they would be prompted to call a neighbour, and then emergency services. KPIs such as falls within a specified period are tracked, and a history of the events for the subscriber’s device. The sensor being out of range or having no movement over a period of time can also trigger a new task instance, while reorienting the sensor to the upright orientation within a few seconds after a fall was detected can cancel the process. Looking at the BPMN for managing events from the sensor, they are using the event objects in standard BPMN to their fullest extent, including both in-line and boundary events, with the device events translating to BPMN signal events. Great example of responsive event handling using BPMN.
Whitestein: Demonstrating Measurable Intelligence in an Enterprise Process Platform
The last demo of the day was Dan Neason of Whitestein also was in the theme of events, but more focused on intelligent agents and measurable intelligence in processes. Their LSPS solution models and executes goal-driven processes, where the system uses previous events to evolve its methods for reaching the goals, predicting outcomes, and recommending alternatives. The scenario used was a mortgage application campaign, where information about applicants is gathered and the success of the campaign determined by the number of completed mortgages; potential fraud cases are detected and recommended actions presented to a user to handle the case. Feedback from the user, in the form of accepting or rejecting recommendations, is used to tune the predictions. In addition to showing standard dashboards of events that have occurred, it can also give a dashboard view of predictions such as how many mortgage applications are expected to fail, including those that may be able to be resolved favorably through some recommended actions. The system is self-learning based on statistical models and domain knowledge, so can detect predefined patterns or completely emergent patterns; it can be applied to provide predictive analytics and goal-seeking behavior across multiple systems, including other BPMS.
Wrapping up this set of demos on intelligent, event-driven processes, we had a keynote from Jim Sinur (formerly of Gartner, now an independent consultant) on goal-directed processes. He covered concepts of hybrid processes, made up of multiple heterogeneous systems and processes that may exhibit both orchestration and collaboration to solve business problems.
Great first set of demos, definitely setting the bar high for tomorrow’s full day of 11 demos, and a good first day. We’re all off to the roof deck for a reception, wine tasting and dinner, so that’s it for blogging for today.
By the way, I realize that we completely forgot to create bpmNEXT bingo cards, although it did take until after 4pm for “ontology” to come up.