Since the advent of the digital twin in 2002, the idea of using computers to model the real world has gained traction outside of aerospace and other hardware-intensive manufacturing industries where it was first introduced.

Now, they're creeping into the workforce. 

At EY, Global CTO Nicola Morini Bianzino is applying digital twins to the entire organization. 

The company is spread across 200 countries and with a workforce of over 350,000 employees, it's hard to know how well the company, as a whole, is doing. 

"What happens is that the metrics on how the business is run are very siloed around the different businesses," he said. "So you can see green [lights] everywhere and everybody's pissed off. Why? Because these [systems] don't connect with each other."

Process twins, as they are known, allow Bianzino to see how the performance of one area of the organization impacts others. He can then make changes to workflows to reduce points of friction and increase efficiency.

"The benefit is that you're able to tailor the [organization's] processes to the reality of things on the ground," he said.

The data and experts agree: the demand for digital twins is catching on, thanks to their ability to improve processes, optimize operations, provide insights where none previously existed, and accelerate innovation, all while having the potential to drive down costs.

The market for digital twins is set for strong growth. In 2020, the market for digital twins was valued at $5.1 billion. By 2026, it is forecast to reach $35.5 billion, a 37% CAGR, according to recent ResarchandMarkets research.

The pandemic drove digital twin adoption, especially in the pharmaceutical and healthcare sectors, the research found. The technology also has a growing role in predictive analytics. 

Alexander Wurm, research analyst covering analytics, data warehouse and integration technology at Nucleus Research, is seeing emerging use cases in the pharmaceutical industry, where digital twins will act as stand-ins for human organs like lungs or a heart. 

These digital twins will use real-time and time-series data, rather than just point-in-time data, to help improve the testing of new drugs like the mRNA-based vaccines used to battle the COVID-19 pandemic.

"Part of the digital twin's appeal is that you can break the normal laws of physics, you can accelerate time, you can do scenarios, you can do a bunch of different things," Wurm said. "A lot of these very sophisticated machine learning and deep learning models, they are very informative, but they're not the focus of what drives business value. Right now, what drives business value is the fast insights that you can act on quickly."

Use cases abound

Home improvement giant Lowe's is developing digital twins to create "intelligent stores" that adapt and evolve in real-time, said Cheryl Friedman, vice president of Lowe’s Innovation Labs. 

"We see a future in which emerging technologies like digital twins, robotics, mixed reality and blockchain rewire how we as Lowe’s understand and optimize what’s happening in our stores," she said. "A retail twin can be the foundation for getting our customers the products and help they need, when and where they need them." 

Lowe's digital twin will combine 3D store maps with e-commerce-style analytics to provide a foundation for training AI models. 

Once in place, these models will guide robots down store aisles that can recognize where store associates are needed or alert them when an item is out of stock. 

Supply chains are starting to use similar approaches. By combining static information – addresses, order date or type of cargo – with real-time telemetry data – locations of containers, weather information or asset condition data– logistics companies get visibility into an entire supply chain in real-time. 

Another element driving adoption is the avalanche of data inundating organizations big and small. 

Companies grapple with data flowing from an ever-expanding army of internet of things (IoT) platforms and devices, network-connected hardware like servers and PCs, apps and applications, and outside sources such as vendors and partners.

Cloud compute cycles and storage create an almost limitless supply of data for digital twins, information needed to model more complex and accurate systems.

"As IoT devices become cheaper and ... real-time analytics become better, that's what will spread [the adoption] of digital twins among different industries," said Wurm.

However, current adoption rates are still low. Less than 5% of businesses use digital twins, he said.

Semantics matter

 A digital twin must incorporate six defining attributes, according to Forrester. These systems must: 

• Represent something in physical world
• Support bi-direction data flows
• Support modeling and predictive analytics
• Provide reporting 
• Send and receive new data and commands as quickly as required
• Be stateful, or remember the last position of its physical counterpart in the event the data connection is lost.

While most people associate digital twins with pretty pictures, many digital twins do not rely on graphical user interfaces of any kind. 

Similar to modern AI-driven analytics platforms, digital twins are often analytical models of real-world systems and processes that digest real-time data and provide reporting designed to augment human decision making. 

For enterprise leaders, this expanded view of what a digital twin is — and what it can do — drives adoption forward, said Bianzino.

"We need another name for it," he said. "Maybe this work on the metaverse will bring another way of defining them."