Use process modelling and digital twins to identify structural bottlenecks and speed up workflows

Transparency into core processes
Process mining is a technology that helps organizations discover, analyse, and improve workflows. A business process leaves digital footprints in the form of event log data. In process mining, an algorithm is applied to the event log to create a living picture of what a business process actually looks like – and not what the company thinks it looks like. By uncovering trends in event log data and using those trends to construct a process model, companies can go on to visualize the workflows occurring within a system and across a system landscape and follow up with targeted initiatives.

The key benefits of process mining (acquiring and analysing data) and process modelling (adding a visual representation) are insights that offer previously unobtainable levels of transparency into core processes.

While gaining insights into how a company runs its business processes is always advisable, there are use cases where process mining and process modelling are particularly relevant, such as M&As, audits, opening up a new office and/or a new market, and so on. Process mining and process modelling can also be used to enhance existing business processes and simulate future change. When you have a digital representation fuelled by event log data, you have a powerful tool to run predictive scenarios and test different strategies for resource allocation, workflow optimization, automation initiatives, and other critical business decisions.

At Deloitte, we have partnered with Celonis and IBM, among others, and use their software to run large-scale process mining and process modelling projects.

Operational optimizations with digital twins
The concept of digital twins is related to process mining and process modelling and possesses some of the same qualities and capabilities for running a more effective business.

A digital twin is a virtual model of a physical asset that 1:1 mimics the behaviour and operation of its physical counterpart. The asset could be a car, an engine, a heart, or anything else that behaves in a certain way because it is a physical ‘thing’. A digital twin is a powerful tool for copying what is happening in the real world and running simulations in a safe, digital environment.

If we dive deeper into the technical aspects of digital twins, they are mathematical models based on the laws of physics. The mathematical model in a digital twin uses a set of equations to describe how core processes work and interact. Those processes express the most essential characteristics of a company’s operating model. The scientific rationale – which can be translated directly into a business potential – is that the more a company understands the present state of a physical system, the faster it can learn, and the better it can run predictions to improve the future state of its physical systems.

Digital twins are used extensively in biophysics, and from there, the technology has spread to the pharmaceutical industry, where it is currently used for personalized medicine and drug manufacturing and will play a vital part in developing psychoactive drugs, wound treatment, arthropathy, etc. Digital Twins are also used in the train, aviation, automobile, space travel, and manufacturing industries.

It is worth mentioning that NVIDIA offers Omniverse, a platform that includes physically accurate simulations and incorporates AI tools for tasks such as generating synthetic data and automating workflows. Its applications span various industries, including gaming, architecture, engineering, and manufacturing.

You need slack and redundancy
At Deloitte, we have designed digital twins for various companies, such as harbours and rail transportation. One of the key learnings from such projects is to look for bottlenecks. Usually, a company can easily define the series of actions in a progression from the start to the finish line that constitutes a business process. But they are often left in the dark when it comes to scrutinizing where, along the way, something goes wrong.

Companies can better understand their maximum output levels for a specific process with a digital twin. For example, how many cases can we handle in this department per week, what is the quantity of ships that can debark from 1 AM-2 AM, how many ampoules can we produce per hour? And so on. Companies usually try to squeeze more output from a business process than possible, thus facing bottlenecks. Theoretically, you may estimate that 100 cars can pass a narrow passage per minute, e.g., going 50 km/h. But that estimation rests on the belief that 100 cars drive bumper-to-dumper through the passage. This scenario is dangerous since a small mistake from a single driver can lead to a pile-up. The estimate is simply too high. You need quite some slack and redundancy in the overall system to buffer against various kinds of disturbances – which is why you should keep at least two seconds between cars.

A piece of advice: To keep your workflow efficient, ensure tasks move together like an assembly line. You have reached your maximum capacity if you are constantly finishing tasks only to replace them with new ones. To handle more, find ways to complete tasks faster or delegate to increase your overall output.