One of the recurring metaverse problems is the tendency to nail the term to the social-media applications, when the concept is actually much broader. A joint report by Nokia and EY says that the “industrial metaverse” is creating “substantial business value” and demonstrates “significant or transformative impact”. This doesn’t surprise me; I’ve advocated a “metaverse of things” (MoT) model that’s a general case of the industrial metaverse being discussed. Further, Nokia and I chatted about their view on this earlier this year, and I noted that they were generating successful MoT implementations in key industries.
The report is fascinating to me because there are places where it tracks what I hear from enterprises very closely. It’s also fascinating because some of those tracking points represent what I believe are weaknesses in the way enterprises are looking at the metaverse, and thus how they might frame their adoption. It has good data, but also data that I think has some issues, and it has one bias that I’m uncomfortable with. Let’s start with my own data.
The good news that the metaverse concept can actually deliver on its promise is offset by the fact that the great majority of companies (over 85%) have little or no understanding of industrial or enterprise applications of the metaverse concept. One thing I use as a measure of whether a company “gets” the MoT model is whether they link a metaverse approach with the broad notion of digital twinning of real-time systems, and this is what’s currently limiting MoT literacy.
The majority of companies I’ve talked with still see the metaverse through a virtual-reality lens, meaning that they actually assume that the consumer-model social-metaverse concept of a virtual world is the foundation of what they like. The report says that 41% of companies interviewed say they have an all-encompassing view of the metaverse, meaning that they see its ultimate state as a business-wide virtual world that their workers, partners, and customers inhabit. A slightly larger number, 44%, see it as the next step in digitization. Both these groups, according to the report, see a virtual-world framework for the metaverse in enterprise/industrial applications, and that’s the bias I have a problem with.
The report divides what I’ve called MoT into three categories, Consumer, Enterprise, and Industrial. I don’t disagree with that division, but the way each category is defined suggests that of the three, only the Industrial one is really linked to the digital-twin concept. I still believe that there’s a symbiosis among the three models and a benefit to addressing them based on a single MoT architecture with tuned features per model, but it’s certainly possible to justify the way they approach this. I think of it as a software architecture challenge as much as a business transformation approach, which the report doesn’t do.
A problem I see with the categorization is that it admits the Enterprise metaverse could be built with collaborative tools, and that the industrial metaverse doesn’t require any digital twinning. This may be why the report says that 58% of users have deployed at least one metaverse mission where my own findings say that only about 8% have done so; a different metaverse definition means different metaverse penetration. Instead, as I’ve noted, the report focuses on virtual worlds and visualization as the common element.
This may be why the report links the metaverse and applications that center on simulations. “Respondents viewed XR [extended reality] onboarding and soft skills training and virtual R&D, testing and prototyping as having the highest potential impact among enterprise and industrial use cases….” My user contacts suggested that these missions were more an exploitation of MoT than the prime justification. A metaverse is a machine representation of a real-world system, though, and one of the things you could do with that is to simulate the behavior of the system under various conditions, even to the point of using the metaverse model to help design what the real-world system would look like in the end.
This approach to MoT offers benefits in that it separates the digital-twin visualization and its rule set from the sensors and controllers that link it to the real world. The model is an XR representation not necessarily in a pure visual way, but in a full behavioral way. That suggests to me that there could be a very significant AI mission in the implementation of MoT. However, the soft commitment to digital twinning implicit in the definitions of the classes of metaverse would weaken this by focusing on a visual representation. That, in turn, may be why the prototyping, training, and testing missions get a high priority in the study; those missions are more likely to be linked to visualization than a mission where the metaverse was tasked with actual control over a real-time system.
The report does, a bit of the way into it, make the point that the “industrial metaverse” includes digital twin technology, which they say is the overlap element with “Industry 4.0”, the kind of mindless media buzzword thing I tend to cringe at. It’s my own view that the evolution of Industry 4.0 concepts are really dependent on digital twinning and MoT, for the simple reason that enhancements to process automation demand a close coupling between the real world and a digital model that can be used to predict, monitor, and control it. However, the report falls prey to the social-metaverse-augmented-reality thing when it says that “Industrial metaverse adds the use of visualization tools”. I think that augmented reality “viewing” options are important for some and perhaps even many MoT missions, but I also think that focusing on AR distorts the opportunity and limits the business case.
One example of this is the use case list the report offers. Yes, virtual reality or augmented reality is very helpful in the use cases listed, but I think that real-time process control and even the control of multi-process systems, including those geographically distributed, is the biggest opportunity. We tend to approach process automation these days by integrating independently automated steps and tasks rather than by organizing the entire process through an automated system. In a digital-twin model, the application of process automation is “fractal” in that if you look at a high-level process model closely, you see it’s made up of an organized set of inner models, constructed in the same way. Yes, you could use digital twinning to automate the welding done on a car body, but you could also use it to manage the other atomic manufacturing tasks, to automate the entire set of tasks into an assembly line, and to integrate that production process with distribution and sales.
I do think that the applications that the report focuses on are likely lower apples to pick. I also think they have more PR buzz, and papers like this are first and foremost marketing material not scientific research. I think it’s correct to say that there’s benefits here that could be tapped readily, and that’s a good thing. My concern is that the focus on this AR/VR-centric, visualization-centric, set of use cases could encourage the creation of software that can’t readily address that fractal business control opportunity, and thus will either cause us to miss that opportunity or significantly delay realizing it, and only then realizing it at a higher cost.
The report’s “Production and delivery” section seems to work to address that, but it still focuses on the visual. “Virtual representations of facilities can also serve as a high-fidelity reporting dashboard for advanced data capture from Industry 4.0 solutions and monitor assets via smart sensors.” This comment is under the heading “Virtual facility automation”, which I would prefer to see called “modeled facility automation” because the term “virtual facility” implies a visual representation of the facility. Is that essential? No. It’s not even always particularly useful, and I think it’s far less important than the use of a digital-twin model to coordinate sensor data and activate control behaviors.
Don’t get me wrong, this is a decent marketing-level report. There’s useful information on user interests and on use cases contained in it. I think there are some issues created by relying on user statements of expected future behavior, the sort of statements I’ve found in my own survey experience to be almost universally unreliable. It does raise most of the benefits and issues we could expect, even if I can’t fully agree on the way it presents them more visually than as a control system. I’d like to see some statements on software architecture, not just because of my own background but because we’re eventually going to have to realize the metaverse though software, and having a single approach and even a single modeling language and toolkit would significantly accelerate our path to the metaverse, reduce costs, and improve benefit realization. Maybe they’ll get to that at some point. What it surely does is position Nokia as a major player in the “industrial metaverse” concept, and perhaps in my MoT notion too. That could prove very valuable for the company and its desire to broaden its base to the enterprise, and it might even show some service providers the direction they should be considering for their future revenue growth.
