Informed business decision making should rely on models of reality that are accurate and that reflect the purpose for which they were built.
The first issue - accuracy - is addressed by relying on the digital footprint of a system to get a grasp of the status quo, thus ensuring that the model is not distilled from what
“people think” happens in a system, or what “the rules” say.
The second issue – alignment of model design choices with a Goal - is addressed by having the goal dictate specific views of the world on which to zoom in: We categorize to get manageable chunks of (labelled / unlabelled) information with close-enough characteristics. We quantify to forecast future parameter values | trends.
And we tell stories to grasp the behaviour of “actors” (people, organisations) in the surrounding environment.
It is now possible to fully automate the extraction of different types of views of reality. Machine learning classification and regression algorithms are the digital version of our in-built mechanisms of categorization and forecasting. Process mining algorithms are the solution to reverse engineer and storify an ordered structure of the end-to-end journeys (sequences of activities that occur within a pre-specified timeframe) from the data, without the need to explicitly specify what to look for.
Process mining can help you:
Save the time needed to get an initial picture of the situation by talking to people on the ground. Process discovery algorithms can get a snapshot of all your process sequences, as they are, quickly.
Figure out the oddities: what sequences of activities stand out in terms of end-to-end time, length, frequency.
Get a checklist of potential opportunities of improvement - business process flow inefficiencies (unusually long end-to-end waiting times, repetitive sequences of steps that incrementally add uncomfortable delays, queues), unexpected behaviour, deviations from expected behaviour.
Compare business process flow and performance indicator values across departments / units.
Compare the observed behaviour against guidelines (for audit purposes).
And easily transfer this knowledge (to new staff / other experts analyzing the system).
Process mining is domain independent, applicable in any domain where flows of processes are of interest: logistics, insurance, retail, banking, manufacturing, healthcare, and more. In banking, for instance, the re-engineering of optimisable processes was shown to decrease time spent on audit by 50%, throughput time by 30% and application processes (for loans, for instance), by 30 minutes per case. In healthcare, process mining was able to achieve 6%–10% increase in speed of throughput, 20% increase in bed utilization, 5% increase in operating room output, and 5%–8% annual savings in operating budget.
To save time and money you must ensure your business runs end-to-end in the best possible way. To own your processes, you must first know your processes. It is only then that you can really check that they are working for you, not against you. And find the best way things should be done.
ThiaperProcess: a process mining solution
Thiaper Systems SRL is a Bucharest-based company. We kicked off the development of ThiaperProcess - our process mining solution - in April 2019, in collaboration with the largest private healthcare provider in Portugal.
You can see ThiaperProcess in action for an open dataset here.
And if you want to know more, come talk to us to see how we can help you! You can find us on the 11-12 November exhibiting in GoTech World 2020.
Comments