Bridging the Gap: Digital Twins and Decision Intelligence
As you might have heard last month on The Economist podcast, digital twins are increasingly important in a data-driven world: they are revolutionizing industries by providing a real-time, digital representation of physical objects, systems, or environments. These virtual models connect to their real-world counterparts through sensors and data streams, enabling continuous monitoring and interaction. Whether it’s a jet engine with its twin in a computer on the ground, or a Formula 1 car, digital twins are used increasingly in high-stakes situations with lots of moving parts.
But modern digital twins are not just static mirrors—they also incorporate artificial intelligence (AI) to predict future states, model potential scenarios, and assess the effects of various actions like maintenance or modification. Whether it’s a factory machine, a city infrastructure, or even an entire supply chain, these modern, AI-powered digital twins offer a comprehensive understanding of complex systems that leads, in turn, to substantial economic benefits.
However, while digital twins offer unparalleled insight, they often come with a hefty price tag. Building and maintaining high-fidelity models that continuously sync with real-world data requires significant resources, both in terms of data infrastructure and computational power.
…while real-time data and AI are invaluable for making decisions, not every decision requires a fully-fledged digital twin.
Digital Twins and Decision Intelligence
Digital twins excel in environments where a high level of detail is necessary to make decisions—think of monitoring a jet engine to make maintenance choices in real time, or managing the dynamic flow of traffic in a smart city. The digital twin enables continuous monitoring, prediction, and action modeling based on massive amounts of data.
But not every decision requires this level of data fidelity. In many cases, decision makers need just enough information to make an informed choice.
By way of example, if you’re use a GPS navigation system in your car, you don’t need real-time meteorological data, traffic flows from hundreds of sensors, or data about the architecture of the buildings you’ll pass. A simplified, actionable view of the situation is enough—maybe just a basic map with traffic updates, plus recommendations on which turn to take next.
Now think of how this translates to a business situation: where decision intelligence provides a “lens” on what otherwise might be a full digital twin model: giving you all the benefits without the cost, because DI focuses you on only the data that’s needed to support your decision.
In your own business, as you imagine how your decisions will impact the future, given knowledge from colleagues and information in various data stores, do you think you can assemble all of this information in your head? For many business decision makers, the answer is “probably not”.
For instance, if you’re managing a new project and need to allocate resources to it, a full-scale digital twin of your organization’s operations may not be necessary. Instead, decision intelligence can analyze project data, workforce performance, and timelines to recommend how best to proceed. In this way, DI allows organizations to experience many of the benefits of digital twins—such as real-time information and AI-driven predictions—without needing the high infrastructure costs associated with high-fidelity models.
So DI is about augmenting human decision making with computer simulation of future events, (read more details about DI for predicting business futures in this guest blog post). DI brings relevant data into focus, delivering the digital twin benefits of real-time monitoring, AI-driven prediction, and action modeling, but without the prohibitive cost. DI can also integrate human knowledge plus structured, unstructured, and incomplete data in order to build the best simulation model possible with limited resources.
DI brings relevant data into focus, delivering the digital twin benefits of real-time monitoring, AI-driven prediction, and action modeling, but without the prohibitive cost.
Integrating physical systems simulation with business and other contexts
Probably the most important way that DI goes beyond digital twins is that it can model “invisible” things in a variety of domains, like profit, morale, and productivity. Digital twins are—in contrast—generally focused on physical behavior of objects like a gear within a jet engine. From this point of view, DI models can use digital twin models as components.
At my company, Quantellia, for example, we did a project during COVID, where we built a digital twin of people walking in a building, along with the Covid particles they emitted and encountered. This was part of a larger model allowing building managers to choose when and how (masks? social distancing?) to allow workers to return to the office.
This model included psychological components (what does it take to convince people to wear a mask?), economic elements (how many people need to be in the office to stay in business?), health knowledge (how many Covid particles will make you sick?), and our digital twin of particle movement into a heterogenous DI model. In this way, it “wrapped” the digital twin component in a larger decision-making context.
Knowledge without data
DI complements digital twins in a second way, which is that DI can go beyond using knowledge that’s captured in the form of data—whether it’s sensor data, operational metrics, or other real-time streams. Human expertise goes far beyond what can be measured by data alone. Throughout history, humans have captured knowledge in various formalisms like if-then rules, written papers based on randomized control trials (RCTs, such as those used in medicine), and even oral tradition. This kind of expertise may not exist in raw data form, but it remains a valuable decision-making ingredient.
For example, a grower might know from experience that, in rainy years, it’s best to plant a particular species of a crop, even if they’ve forgotten the reason why. DI allows for the integration of these if-then rules, along with other forms of expert knowledge, ensuring that decisions are informed not just by numbers but also by the wisdom and intuition that have guided humans for centuries. In scenarios where data is scarce or nonexistent, DI can draw on this expert knowledge to provide actionable insights that digital twins might miss entirely.
DI models also represent a way to use this kind of expertise in novel scenarios, by combining small elements of knowledge (“this crop is good in rainy years”) into entirely new situations, such as might be encountered with climate change.
Another example: a supply-chain expert might have once advocated for just-in-time inventory systems, but after the pandemic, with supply chain disruptions being more common, that strategy may no longer be feasible. Instead of discarding the expertise altogether, DI can integrate the expert’s understanding of why just-in-time worked in the past and use it to navigate the present by suggesting alternate approaches.
In scenarios where data is scarce or nonexistent, DI can draw on … expert knowledge to provide actionable insights that digital twins might miss entirely.
DI and digital twins: a robust partnership
DI doesn’t just mimic the real world through data. It enriches decision making by embracing human knowledge that’s often overlooked, and by keeping expert insights relevant, even as circumstances evolve. DI helps organizations maintain an agile and responsive approach, making it an ideal complement to digital twins, and one that’s a lower-cost alternative in many situations.
Together, digital twins and DI offer a robust framework for decisions, blending the best of real-time data with the depth of human expertise.
Sound valuable? At Quantellia we’ve been building DI and digital twin models for over 15 years. We work with small and large organizations alike, providing consulting plus software services and solutions in DI, ML, and GenAI across industries like transportation, health care, financial services, government budgeting, education, and many more. Check out our offerings at www.quantellia.com or drop me a line any time if we can be of service.
