With all these RPA / Workflow / ML (still emergent) companies starting to merge their capabilities, we will begin to see self-assembling patterns of services. The notion of loosely coupled services that self assemble into larger flows will require model-driven analytics (not the log-based reporting which is after the fact, but companies are going to need the ability to anticipate what might happen. This is a bit scary since most compliance-based operations are all hard-wired ( = zero adaptability). This 1980’s model will no longer make sense, so companies are going to be forced to instantiate their policies into models (not simple rule sets that are impossible to navigate let alone understand collectively). All of this suggests the need to create a sub-assemblies of emergent processes within a given operational domain (suggesting the need for the domain models that I’ve socialized in the last few years) which can provide a measure of predictability and thus support patterns of aligned policy. Suddenly, all these BI and Big Data analytics no longer seem relevant. They are going to struggle to ingest multiple sets of patterns into their reporting analytics and algorithms. This may give rise to a new set of model-driven analytic tools.
Further, very few companies have the ability to define, build, and maintain models so they are going to need tools that help them build models and traverse them in a more visually inviting manner. The growing importance of true knowledge management tools (Franz, PoolParty, etc.) will start to play a larger role in organizations. Companies will shift their investment in algorithms into knowledge engineers, who can model knowledge graphs under the control of business so they can simplify the algorithmic landscape to leverage the patterns of knowledge that are instantiated in proper containers (knowledge graphs NOT algorithms). Then I believe we may start to see the need for model-driven analytic tools that the business can use with the help of their knowledge engineers to build on-demand dashboards that can provide a real-time simulation of their operational and (eventually) strategic landscapes.
Comments