The storyline of Sliding Doors, a British-American romantic drama, splits into two after its heroine Helen Quilley gets fired from her PR job. In one of the parallel universes that opens, Helen catches the train, while in the other she misses it. Two completely different lines of events unfold, where the relative inconvenience of missing a train results in a considerably better long-term outcome.
It may sound corny, but the sliding door of the film is a good illustration of what companies in two minds about embarking on a digital journey need. There would certainly be less uncertainty about whether to jump on the digitalisation train or not if decision-makers could enter parallel universes and see the costs and benefits of chosen alternatives as well as the risks they incur. And this is exactly what predictive simulation and digital twins enable them to do.
The concept of digital twins – the concept of dynamically evolving, real-time replicas of assets, systems and processes – is already fully-fledged. But what really drives its advancement today is the industrial internet of things (IIoT) and the extensive use of sensors that provide a barrage of real-time data.
Some advantages of the digital twin technology manifest themselves very clearly: it removes silos, identifies operational bottlenecks and quickens innovation cycles. In manufacturing it helps identify equipment problem spots and can reduce downtime by remote maintenance, or by alerting operators of systems to technical problems before they occur.
But predictive simulation – and the digital twins it produces – can play an equally important role in de-risking technological investment. In times of big technological shifts such as the current one, when cutting-edge technologies abound, using a dynamic model of a system or process and testing the outcomes and return on investment virtually before making a commitment can contribute significantly to the success of a transformation project. With digital twins, decision makers will have the opportunity to simulate selected technological scenarios, analyse outcomes and compare them against each other before deciding on the best course of action.
Once the ideal transformation path is identified, there are still a number of pitfalls to avoid during implementation, some of which are AI-related. Whether your business is an incumbent or a digital native, dips its toe into digitalisation with a pilot project or opts for a complete digital overhaul, the chances are you will deploy AI for purposes as diverse as data analytics, fraud detection, identity management, cyber-security or cognitive robotics.
However, as GE Digital’s John Renick points out in an interview, AI projects tend to underwhelm and underdeliver. AI is fuelled by data and will be only as good as the quality of the data you put into it. Internal historical data will not get you too far, and to receive the right answers from an AI system, data experts need to understand current and expected behaviour as well. And this is where digital twins can be of invaluable help as additional, external data sources.
If statistics and trend analyses are anything to go buy, digitalisation is gaining traction at breakneck speed. Missing the train – unlike in Sliding Doors – is no recipe for survival. But boarding a random train – that is, investing in digital technology without ensuring the outcome will provide the business with a lasting competitive edge – is no guarantee for success either. That is why tools such as digital twins, that enable informed investment allocation decisions, are more relevant than ever.