Just finishing up some notes from my trip to SAP TechEd && d-code last week with the latest on their Operational Process Intelligence product, which can pull events and data from multiple systems – including SAP’s ERP and other core enterprise systems as well as SAP BPM – and provides real-time analytics via their HANA in-memory database. I attended a session on this, then had an individual briefing later to round things out.
Big processes are becoming a thing, and if you have big processes (that is, processes that span multiple systems, and consume/emit big data and high volume from a variety of sources), you need to have operational intelligence integrated into those processes. SAP is addressing this with their SAP Operational Process Intelligence, or what they see as a GPS for your business: a holistic view of where you are relative to your goals, the obstacles in your path, and the best way to reach your goals. It’s not just about what has happened already (traditional business intelligence), but what is happening right now (real-time analytics), what is going to happen (predictive analytics) and the ability to adjust the business process to accommodate the changing environment (sense and respond). Furthermore, it includes data and events from multiple systems, hence needs to provide scope beyond any one system’s analytics; narrow scope has been a major shortcoming of BPMS-based analytics in the past.
In a breakout session, Thomas Volmering and Harsh Jegadeesan gave an update and demo on the latest in their OPInt product. There are some new visualization features since I last saw it, plus the ability to do more with guided tasks including kicking off other processes, and trigger alerts based on KPIs. Their demo is based on a real logistics hub operation, which combines a wide variety of people, processes and systems, with the added complexity of physical goods movement.
Although rules have always been a part of their product suite, BRM is being highlighted as a more active participant in detecting conditions, then making predictions and recommendations, leveraging the ability to run rules directly in HANA: putting real-time guardrails around a business process or scenario. They also use rules to instantiate processes in BPM, such as for exception handling. This closer integration of rules is new since I last saw OPInt back at SAPPHIRE, and clearly elevates this from an analytics application to an operational intelligence platform that can sense and respond to events. Since SAP BPM has been able to use HANA as a database platform for at least a year, I assume that we will eventually see some BPM functionality (besides simple queuing) pushed down into HANA, as they have done with BRM, allowing for more predictive behavior and analytics-dependent functions such as work management to be built into BPM processes. As it is, hosting BPM on HANA allows the real-time data to be integrated directly into any other analytics, including OPInt.
OPInt provides ad hoc task management using a modern collaborative UI to define actions, tasks and participants; this is providing the primary “case management” capability now, although it’s really a somewhat simpler collaborative task management. With HANA behind the scenes, however, there is the opportunity for SAP to take this further down the road towards full case management, although the separation of this from their BPM platform may not prove to be a good thing for all of the hybrid structured/unstructured processes out there.
The creation of the underlying models looks similar to what I’ve been seeing from them for a while: the business scenario is defined as a graphical flow model (or imported from a process in Business Suite), organized into phases and milestones that will frame the visualization, and connected to the data sources; but now the rules can be identified directly on the process elements. The dashboard is automatically created, although it can be customized. In a new view (currently still in the lab), you will also be able to see the underlying process model with overlaid analytics, e.g., cost data; this seems like a perfect opportunity for a process mining/discovery visualization, although that’s more of a tool for an analyst than whoever might be monitoring a process in real-time.