More and more companies today are seeking to add digital twins to their operations. Put simply, a digital twin is a digital model of a physical object, system or process. They can be used to monitor, analyze and simulate their physical counterparts. But implementing digital twins can be tricky for companies that are unfamiliar with this technology. What must organizations do to avoid issues and succeed with digital twins.

By Teresa Tung, Marc Carrel-Billiard, Matthew Thomas, and Sarat Maitin, Accenture

<<< Start >>>

Artificial Intelligence (AI) dramatically enhances the capabilities of these digital doubles. | Image: Accenture

<<< End >>>

Many organizations have at least begun to experiment with the use of digital twins. These are digital representations of physical things. Using twins for testing during the validation phase of a design process is a popular starting point. But these implementations enhance or improve just one piece of a larger process. They barely scratch the surface of what the digital twin can accomplish.

In fact, done right, it will help to usher in the future of product development. That is, a total process transformation to boost efficiencies, speed up development and deliver better design options and generate new revenue streams.

Digital twins must be deployed with AI. This enables companies to achieve new, intelligent and resilient product design. The physical manifestation of a product can be pushed closer to the supply chain and the customer. Design and manufacturing cycles shrink from years to weeks. And creativity is unleashed with unprecedented force.

Multiple goals can be realized through the next evolution of product development. These include customer-led design and ecosystem-enabled innovation. Likewise, agile product development can also be achieved. This helps companies to quickly respond new challenges. For example, changing preferences and regulations and supply chain issues.


<<< Start >>>

Our vision combines AI and automation with data and models within the twin to reimagine and reshape the entire product development process.

<<< End >>>


This vision also results in products that are never “finished”. Instead, they evolve continuously through updates. This occurs even after they are in customers’ hands. A fully-fledged digital twin and AI solution supports connections in both directions. Released products can capture information on real-world usage and customer needs. This data can be immediately used to design new functionality.

The digital twin in combination with AI powers companies to reimagine how they develop products. It’s possible for even the simplest “static” products to become “living” products. And it ensures companies can deliver an evolving experience for their customers.


A customer-led model

Traditional product life cycles are driven by human engineers. And they are primarily influenced by technology advancements. AI, however, is transforming this relationship. Customer data on how they use products can now be applied directly during the design process. This is a complete inversion of the traditional R&D-to-manufacturing-to-marketing model. With this approach, “lot size one” manufacturing becomes possible. It enables highly personalized, contextualized and even individual-use products.


<<< Start >>>

Get to grips with AI and IoT

Find out how Accenture leveraged AI and the sensor technologies of the Internet of Things (IoT) to help Ducati Corse optimize their MotoGP racing bike configuration to suit every possible race.


<<< End >>>


This same is possible for consumer products. Traditionally, sporadic and delayed customer feedback was saved for future development efforts. Now, the customer’s voice is present from the start. And it continues through manufacturing and into support as continuous monitoring directs the product to update itself. As a result, companies can create richer and more individualized experiences. The type of experiences that have previously been limited to the most expensive luxury goods.


Agile development

To truly transform product development, everyone in the process needs access to a continuously updated single source of truth. But it must be tailored for their own responsibilities. The digital twin done right supports this kind of visibility, e.g. cost and lead time of parts, materials, and manufacturing skillsets. It supports “what if” scenarios that show how a decision will affect other aspects of development. And it also supports the scale needed to serve each customer.

What’s more, collaboration isn’t limited to human teams. AI assistants can work alongside human designers to streamline the data captured by the twin. This helps to create a seamless, iterative process. For instance, an AI-powered design assistant for automotive interiors can walk a designer step-by-step through a design spec. Likewise, automated tax preparation software can guide users through the filing process. And the software will generate increasingly specific questions based on the user’s past answers.


<<< Start >>>

mySpeaker: the future of products

Accenture leveraged this new approach to build the next generation of mobile, smart speaker. Every stage of product development, from initial design, to engineering, to manufacturing, is AI-assisted. Extended reality supports collaboration across stages and from disparate spaces. And the end user becomes a co-creator of the product’s evolution. Consumers share both direct and indirect feedback for future features and product evolutions just by using the product, and designers get a clear view of potential supply chain impacts on manufacturing early in the design process, allowing for quick pivots to alternate parts and designs that keep development on track. A disjointed and linear process with a lengthy time to market becomes collaborative, continuous, and creative—not just end-to-end, but evolutionary.

<<< End >>>


Many product design principles and procedures can be embedded into a digital format. An infinite number of possible products can then be generated from this format. Such software-like representations enable the use of these designs in AI-driven automated tools. Digital twins support combinations of designs, products, and applications. This can help designers explore a variety of different brand identity options. And it allows for more creative exploration in less time.


An ecosystem of innovation

Today, R&D teams and designers act more like platform developers. They manage communities of suppliers, ecosystem partners and end users. These players are all responsible for different parts of the product design process. In fact, multiple partners collaborate to combine mechanical, electrical, and software components. Together, they envision and test the functionality of the integrated product. To succeed, they need tools that operate across different deployments. Such tools must also be able to manage exponential quantities of data.

Accenture uses digital twins with AI to enable tools capable of handling feedback on this scale. Such tools support ecosystem management by identifying the most critical issues during product development. And they provide proactive input and assistance to the ecosystem’s members. This allows for powerful collaboration on development efforts far beyond what we think of as traditional “products.”


<<< Start >>>

Digital twins facilitate dialogue between designers, clients, marketing, product managers and manufacturers. | Image: Accenture

<<< End >>>

For example, we’re using AI to automatically propose changes to a bill of materials for 5G network site planning based on local market demands. This relies on the digital twin to capture context-specific knowledge (geography-specific information, for example) about the appropriate configurations of the components in the bill of materials. The digital twin also indicates how components should be configured in combination with others.

When introducing a new capability like 5G, we capture how to model the necessary materials for a site deployment. This includes everything from a radio to the screws and brackets needed to hold things in place. The model also includes the cost, the impact on transmissions, suppliers and service. And it will generally feature enough detail needed for the end user to configure.

The AI solution can then adapt based on future changes. For example, if an OEM recommends an upgrade to a component, the AI-powered solution can determine if it should be used in that configuration. Or it may flag it as being incompatible with the other parts in the bill of materials. The digital twins facilitate dialogue between designers, clients, marketing, product managers and manufacturers. It enables these players to communicate what is possible, new or desirable.


The building blocks of scale

We’re scaling digital twin technologies to break new ground. It helps us determine users’ desires for new products and features. And it guides us as we translate this information into actionable insights for manufacturers, designers and the end consumer.

Companies will begin their journeys from different starting points. The transformation toward the future of product development is an evolution. It’s not achieved with the flip of the switch. We see AI and digital twin capabilities as a two-dimensional journey, with increasing value maturing toward a vision of completely automated design. How far companies will get on their journey—and how quickly—depend on two dimensions: data and AI impact.


<<< Start >>>

Unlock immediate value and build a roadmap toward complete transformation.

<<< End >>>


As companies progress among the stages of AI maturity, they will move from intelligent analytics all the way up to the full capabilities of an AI and digital twin solution: full process automation.

Intelligent analytics provide a quick way to get started. Specifically, they enable companies to augment the product development process with dashboards and control towers. Moving forward, the twin can support stakeholder synchronization. This allows for different stakeholders across different processes to see the same data as one another. The views of the data may be stakeholder-specific. But they enable collaboration from the same base of truth.

The next level of maturity supports stakeholder coaching. That is, intelligent assistants' coach human stakeholders to complete activities. At the height of both data and application maturity, an organization can reach full process automation. The digital twin automates individual tasks or a process so that it requires little to no human intervention. At the most basic, process automation automates rote tasks based on pattern matching and rules that could be trained to almost anyone and different people would have the same outcome. At the most mature, cognitive automation applies to tasks that require higher level reasoning. These problems do not have single answers but rather multiple options that could be debated. AI-systems need to make choices and apply judgements by reasoning over captured expert knowledge.

When paired with increasingly sophisticated intelligence, different levels of data can unlock different capabilities. Over time, organizations can invest in capturing more data or in building more intelligence for the data they have. Both investments can unlock value. And both will be required to ultimately create more intelligent, automated and user-driven product development.

Designing the future of companies

Digital twins can make the real world machine-readable and machine-controllable. That’s why combining twins with AI is such a powerful concept. We can automate routine tasks and enable continuous and enhanced collaboration. It also helps companies to focus their human experts’ time on new value streams and moving to market with new, groundbreaking products at speed. Design, prototyping, testing, and validation can all be done in a virtual space. This permits more experimentation. And every change can be rapidly evaluated for its impacts before the product is manufactured.

The digital twin done right enables companies to reimagine product development, but the true transformation is even larger. Ultimately, building these capabilities will determine the difference between companies locked in the past—unable to keep pace in today’s marketplace—and the living, evolving companies of the future.


<<< Start >>>

<<< End >>>


About the authors


Teresa Tung, Marc Carrel-Billiard, Matthew Thomas, and Sarat Maitin

Teresa is Managing Director in Accenture Labs responsible for its global Systems & Platforms R&D group whose projects include semantic modeling, edge analytics, and robotics.​ Meet her on LinkedIn

Marc is Global Senior Managing Director for Technology Innovation. In his role Marc leads Accenture Labs, the company's dedicated R&D organization, Accenture Liquid Studios, Accenture Ventures, and Accenture Incubation practices such as Blockchain and Extended Reality. In his role, he also directs Accenture’s annual Technology Vision research, which looks at the future of enterprise technology. Get in touch with him via LinkedIn

Matthew is Managing Director at Accenture within Industry X and leads the Engineering X practice for North America and Siemens Alliance Partnership. Start a conversation with him on LinkedIn

Sarat is Managing Director at Accenture and leads Strategy & Consulting Industrial in Central Europe. Prior to that he was leading the Industry X and Supply Chain & Operations Consulting Practice for Products in Europe. Contact him via LinkedIn


<<< Start >>>


The Authors of the article thank the following for providing valuable insights:

Alex Kass, Mike Kuniavsky, Nik Martelaro and Michelle Sipics from Accenture Labs, and Daniela Mitterbuchner from Industry X, for their contributions to this research and publication.

<<< End >>>

Subscribe to Accenture's The industry X magazine Blog Subscribe to Accenture's The industry X magazine Blog