Unlocking the power of digital twins
Digital twins have come a long way since the concept was first described in an early 2000s systems engineering book. The basic premise remains the same, a physical place or system, for example a manufacturing line, power plant or warehouse, is closely monitored using some kind of connected camera or other sensors. The data captured is used to create an exact digital representation of the system, which can then be used to simulate scenarios digitally before actions are taken in the real world. The benefit of this is the ability to explore the impact of changes on the system without having to actually make them, increasing the range of options you can try, avoiding the disruption as you test them, and allowing faster iteration of the systems making those managing them more agile.
What has evolved significantly are the technologies used to deploy this concept and the domains in which it is relevant, moving well beyond early and low impact use cases powering simple dashboards for IoT enabled systems in manufacturing.
Adding Cloud, AI, and Simulation
Since 2019 we have seen ideas around digital twins become ever more prevalent in a wide range of enterprises, and several enabling technologies have revolutionised the ways in which you can work with them.
The first significant advancement was the migration of digital twins to the cloud, enabling accessibility to unlimited compute resources for interacting with the twin. Combined with AI and simulation, this has exponentially expanded the capabilities of digital twins as a decision making tool. Organisations can harness these flexible computing resources to take a digital twin model and change parameters through conducting millions of complex simulations in parallel, to find the best possible solutions for the physical system. Examples of companies leveraging this approach include Atomico investment Spacemaker (now part of Autodesk) and others like Continuum Industries.
Similarly, the application of other machine learning algorithms enables intelligent optimisation of physical systems, as Atomico investment Oden demonstrates in the manufacturing sector where they use AI to optimise the control variables on production lines to improve throughput or reduce downtime, or investment CloudNC shows, whose digital representation of each component and the manufacturing environment allows them to use AI to optimise cutting paths before physical production, reducing time to machine by 80%.
Where these systems are highly connected and controllable, you can even create closed loop systems where the digital twin directly controls the physical system for continuous improvement.
Introducing Dexory: Harnessing the Power of Digital Twins
A few weeks ago we were thrilled to announce Atomico's latest investment in Dexory, a fantastic company revolutionising warehouse operations through their innovative digital twin platform. Dexory creates a digital twin of a warehouses' physical space and operations using their proprietary data capture hardware, which enables them to create a perfect representation of a non-connected system, a hard problem in the digital twin domain.
Their software platform then enables the twin to be interacted with through powerful analytics related to areas such as:
Inventory tracking including finding missing items or detecting items with certain hazard or customs classifications which could have significant fines associated with any issues;
Space utilisation optimization where better positioning of items based on a clear understanding of item level used and available volumes enables you to increase commercial utilisation of a warehouse;
As well as other areas with significant ROI cases through revenue optimisation or cost saving potential.
Dexory's platform is a great example of how simulation and AI can empower business owners to optimise their operations, enhance productivity, and achieve greater commercial success. For example customer Maersk saved dozens of hours a week in tracking and solving inventory issues through having continuous visibility across their site via the digital twin model, and Menzies saved tens of hours per week automating bond checks through the digital twin allowing warehouse employees to concentrate on more complex and less repetitive tasks.
The combination of a strong technology solution for creating an accurate digital twin model in a “hard to capture” environment, and software which already enables customers to realise significant commercial value made Dexory stand out from other digital twin opportunities for us, and was a big driver of our excitement to invest and partner with them.
Abstraction - Enterprise Digital Twin
In 2019 Gartner coined the concept of the "Digital Twin of Organization" now commonly referred to as the Enterprise Digital Twin. This idea made the concept more abstract, by considering an entire business as the ‘physical system’. By combining this with process mining and incorporating simulation and AI, strategic decisions could be made to enhance the overall performance and operational efficiency of an organisation.
I have been fascinated by the potential of the Enterprise Digital Twin and have come across a small handful of startups working on this idea, developing innovative products that possess significant potential as they mature. Moreover, recent demonstrations of AutoGPT have shown great promise for the possibilities of leveraging LLM technology to unlock even greater abilities for digital twins.
The Time is Now
We believe the time is now for digital twins, and see Dexory as a testament to the immense potential of the technology and its ability to transform industries. We are extremely excited to join forces with Andrei, Oana, Adrian, and the entire Dexory team on this journey, and are actively looking to speak to other individuals and teams involved in creating products in this exciting and transformative area, particularly those focussed on the concept of Enterprise Digital Twins.
