From manufacturing to medicine: How digital twins can unlock new industry advantages
With technology advances, virtual replicas can deliver real gains in more dynamic environments. Explore how digital twins can unlock new value and give industries a competitive edge.
The combination of more accessible reality capture data (think data captured through drones, sensors, and cameras) and sophisticated AI techniques appears to be spurring a new frontier of digital twin use cases. These new opportunities for digital twins—virtual replicas of physical processes or systems—can empower businesses to simulate more complex, open environments that were often previously too difficult to model in a virtual environment. Now, technological innovations with digital twins can create new opportunities for value creation, including uncovering more efficient capital allocation and more effective strategy decisions that can enable new growth opportunities, executed in a cost-effective and worker-friendly manner.
As these capabilities continue to expand, both in terms of sophistication and accessibility, the scope of digital twin opportunities is beginning to touch every industry. A Deloitte analysis that includes interviews with digital twin specialists and an exploration of digital twin use cases highlights how organizations can gain an industry advantage by applying innovative approaches to address some common historical challenges to implementing digital twins in dynamic physical environments.
The evolution of digital twins: Open environments
For decades, manufacturers and engineers have harnessed the power of digital twins to optimize operational performance. For example, racing teams create digital twins of parts, like tires or brakes, to simulate and test performance thresholds before the car ever hits the track. Similarly, smart manufacturers embed sensors across the factory floors to dynamically simulate production and uncover new opportunities to improve efficiencies and reduce costs.
Until recently, many digital twin use cases have been anchored in the realms of manufacturing and engineering due to practical limitations in capturing and processing data required for more complex and unpredictable environments.
However, there looks to be growing momentum to expand the scope of digital twin applications. In our previous article, “New uses for digital twins in the race to navigate an uncertain future,” we explored a variety of new use cases for combining digital twins with simulation methods to better navigate uncertain environments (like simulating a potential merger between two companies or the marketing effects of a new product launch). New forms of data capture and processing can help to address issues around data availability and systems integration.
This momentum appears to be growing beyond just a handful of use cases. According to market estimates, the global digital twin market size is forecasted to increase from nearly US$13 billion in 2023 to US$259 billion by 2032.1 And as noted in Deloitte’s 2024 Future of the Digital Customer Experience survey, many of the technologies that underpin digital twins are becoming more commonplace across industries. For example, 84% of respondents indicate they currently use cloud computing; 72% use Internet of Things (IoT) sensors, devices, and platforms; and 26% deploy extended reality and augmented and virtual reality capabilities (with an additional 26% indicating they plan to deploy these capabilities over the next three years).
The dynamic duo: Digital twins and simulations
Digital twins and simulations are two sides of the same coin, working together to create a tool for understanding and optimizing complex systems. Think of a digital twin as a highly detailed blueprint—a digital replica of a physical or an abstract system. This replica, built through robust mapping of the real-world counterpart, provides a foundation for running sophisticated simulations. These simulations, employing methods like Monte Carlo,2 agent-based modeling,3 or discrete event simulation,4 can allow organizations to test different scenarios in dynamic and comprehensive ways, revealing potential outcomes, identifying vulnerabilities, and ultimately informing better decision-making. While the digital twin provides the detailed representation, it’s the simulation that unlocks its true value, transforming a static replica into a dynamic tool for exploration and optimization.
Reaching the unreachable: Bringing new spatial data to industry solutions
Even in contained environments, like manufacturing plants, capturing data from multiple sources—and centralizing that data—can be difficult. Those data capture and centralization challenges can be more complicated in open spaces, like repairing a skyscraper or bridge located high above rugged terrain, or conversely, in areas so small and dynamic, like parts of the human body, where embedding sensors may not be appropriate.
Nonetheless, addressing these types of problems should include an approach to data capture that better mimics the physical realm. For this reason, some organizations are using various technologies to harness the underlying spatial data that informs these models. Spatial computing brings together operational data (for example, real-time IoT data), business data (for example, enterprise resource planning data), reality-capture data (for example, lidar, or light detection and ranging), geospatial data including temporal data, and three-dimensional data (for example, computer-aided design) to better represent the physical world in a virtual environment.5
While tying these data sources together can be challenging, there are innovative examples of industry players bringing these concepts to life.
Government and public services: Building digital bridges for Cincinnati
Cincinnati’s critical infrastructure includes bridges that are over 150 years old or carry over 100,000 vehicles daily. Inspectors spend months on each identifying emerging issues like cracks and rust to be fixed before any catastrophe.
To drastically improve speed and accuracy, Cincinnati’s mayor evaluated drones enhanced with OptoAI™ real-time onboard analysis. He found that it reduced the task from months to minutes by pinpointing issues, empowering inspectors to rapidly review them on the web or in virtual reality, and generating work orders.6
Life science and health care: Treating seizures through personalized digital twins
While seizures have some known stimuli, such as lighting and acoustics, many occur in seemingly unpredictable ways. In an attempt to help patients better manage these unpredictable events, doctors are experimenting with digital twins as a means to better treat epilepsy patients.
In addition to incorporating brain imagery and data, a Bluetooth-enabled epilepsy device is directly implanted into the brain to allow bidirectional data flow for more dynamic treatment adjustments. “This technology, embedding electrodes that interact through IoT, can allow for real-time monitoring and predictive simulations,” suggests one digital twin architect.7
The architect goes on to explain, “The use of AI and machine learning in these implants means they continuously adapt and optimize treatment based on real-time data, embodying a future where technology and healthcare converge to offer personalized, data-driven solutions.”
Filling the data gaps: Where synthetic data is opening new vistas for simulation
While technology is making spatial data more accessible, sometimes there simply isn’t enough data to capture in the first place. In these cases, it can be difficult—or seemingly impossible—to build a robust digital twin and simulation on sparse data.
To circumvent this very real challenge, organizations are experimenting with synthetic AI data to help fill the information gaps. Consider the following examples.
Energy, resources, and industrials: Generating synthetic data to detect vulnerabilities in the grid
A large power and utilities provider uses computer vision algorithms to identify defects in the physical power grid. But there’s a catch: Image data of key defects is often difficult to find and capture on a sprawling grid, which can make it impossible to train the algorithm with the necessary amount of data to properly identify crucial defects.
To fill this data gap, the provider created over 2,000 images of 3D synthetic data to better inform the algorithm. Using this 3D modeling capability, the provider was able to simulate any type of defect and the scenarios that most likely required remediation. These synthetically fueled models improved defect detection by 67%, leading to reduced asset downtime and higher customer satisfaction.8
Technology, media, and telecommunications: Paving the way to autonomous vehicles
In the race to develop autonomous vehicles, massive amounts of data will be required to inform the complex models that will inevitably drive the cars. In these cases, some organizations are turning to highly detailed synthetic data to better inform their simulated models for training data.
For instance, technology companies Nvidia and Uber recently announced a partnership that will combine Nvidia’s AI infrastructure for training and refining AI models with Uber’s extensive trip data.9 As Dara Khosrowshahi, chief executive officer of Uber, explains in an interview with The Tech Portal, “Generative AI will power the future of mobility, requiring both rich data and very powerful compute. By working with NVIDIA, we are confident that we can help supercharge the timeline for safe and scalable autonomous driving solutions for the industry.”
Addressing the unpredictability of human decision making: The semi-open system
Unlike a piece of equipment, human decisions are much less predictable and, importantly, not always appropriate to track for numerous privacy reasons. In these cases, it can be difficult to create a realistic digital twin of a physical space that is modeled on human actions.
For these reasons, some organizations are using digital twins to either interact with customers and work with them to determine optimal designs of the physical space or deploy digital twin technologies to refine the operational flow of inventory and, inevitably, make customer navigation more intuitive. Consider the following two industry examples.
Consumer and retail: Optimizing operations to clear the way for strategic thinking
Quick service restaurants (QSRs) are under pressure to provide quality food in an accurate and timely manner and at a reasonable cost to customers. To do this effectively, they need to design world-class operational processes enabled through innovative technologies. One North American QSR chain leans on digital twins to drive its operational process design. Through the digital twin, the QSR chain replicates the entire end-to-end fulfillment process to identify, assess, and improve operational opportunities and, importantly, evaluate the impact of different scenarios to determine the most effective path forward.10
Specifically, the chain developed a 3D visualization of the restaurant and simulates different scenarios relating to customer traffic, ordering, and payment and fulfillment. These scenarios contribute to operational, digital, marketing, and layout decisions.
Financial services: Preparing for acquisition
When North American bank, BMO, acquired over 500 locations in the acquisition of Bank of the West, it knew this also meant undergoing the resource-intensive process of redesigning and rebranding each of those new locations.
Rather than sending individuals to each bank to survey the new locations, BMO employed digital twin developer, Matterport, to use its 3D capture technology to help develop virtual replicas of the 500 branches. From these digital twins, BMO simulated branch layouts, encompassing everything from branding to ATM locations to signage in order to more effectively redesign each new space, saving hundreds of thousands in travel costs and over 6,000 hours of survey work.11
Bringing the digital twin advantage to your industry
As data and technology continue to expand, new opportunities for digital twins are expected to arise across every industry. But even in the face of technological innovation, real challenges still exist in creating digital twins that accurately simulate complex, open environments. In an effort to head off those challenges from the onset, industry leaders can consider the following moves to help them design a more dynamic future.
- Identify the value gaps. Value can take many forms, from improving the efficiency of decision making to innovating entirely new sources of value. For leaders weighing where to invest in new digital twin opportunities, consider where the organization needs to more effectively capture—or create—value. Identifying these value gaps can inform whether these technologies should be deployed within the research and design process to increase the speed of innovation or, alternatively, within the operational infrastructure to capture new forms of efficiency.
- Look beyond the organization to build the talent base. Organizations can look beyond their own boundaries to establish the necessary technical and talent infrastructure to support new uses for digital twins. Partnerships, particularly with leading universities, can be instrumental in attracting fresh talent and securing a pipeline of future leaders familiar with the technology.12
- Design for multimodal data. Especially for use cases involving spatial computing, it can be a challenge to build a robust multimodal data model that connects data and workflows across the enterprise. For this reason, leaders can benefit from designing and investing in the proper data infrastructure from the outset to ensure the organization is ready to implement their digital twin use case, even in a complex physical environment.
Across many industries, organizations that can push their digital twin use cases into new realms of the business have the opportunity to build a more resilient, adaptable, and forward-thinking enterprise to navigate a dynamic modern business landscape.
Methodology
Deloitte interviewed 12 senior leaders from organizations across industries, each with US$500 million or more in annual revenue, between May and June 2024. We spoke to these experts (including digital twin architects, strategists, consultants, and business product managers) to understand how digital twins, combined with strategic simulations, are adding value to their organization’s strategy and helping them manage uncertainty. We carried out a thematic analysis of the interview data and reviewed relevant literature to identify the different benefits of deploying digital twins, even in the most ambiguous of environments.
Continue the conversation
Mike Segala
Frances Yu
by
Frances Yu
Brian Campbell
Timothy Murphy
Cover image by: Jaime Austin; Getty Images
Visit the Deloitte Center for Integrated Research
Access more insights on some of the most complex issues facing businesses today.