Podcast: How modern networks are helping industrial AI deliver on the promise

Rajeev Shah, who is co-founder and CEO of Celona, joins SI’s Scott Achelpohl for an episode about how a first step in plant upgrades should be modernizing your network to meet the demands of AI-driven operations and applications. Rajeev is a recognized thought leader on enterprise connectivity and industrial intelligence.

Podcast: Texas Pride Trailers a prime example of digital transformation success

With AI becoming so omnipresent, especially now in manufacturing applications, we at Smart Industry have taken it upon ourselves in 2025 to not so much admire AI as the “shiny, new thing” but really try to dive deeply into use cases for this technology and recognize the hurdles to adoption.

Two recent studies by Hexagon show the disruptive influence that AI is having and why it’s so important for manufacturers to be thinking about AI. It’s well beyond hype now.

In a survey of 500 manufacturing employees—C-suite executives through senior and mid-level managers to entry-level employees—65% in the Americas-focused poll ranked easy integration with current systems (including OT) at the same level of importance as potential cost savings that AI could bring.

It’s rare for companies, always bottom-line oriented, to be thinking of something more than what some tech can do for the bottom line, but AI is having that effect.

In another poll of 1,051 employees from facilities worldwide, 91% expected AI investment to deliver quality improvement and to help solve what that study called an “agility emergency” among manufacturers, especially now that they are under pressure to move toward reshoring and, in the interim, make their supply chains and materials sources nimbler in the Trump tariff economy.

Obviously, part of all this is industrial enterprises’ ability to overcome “legacy” infrastructure challenges—particularly with their OT—and we invited Rajeev to join us to offer more than a few pointers on how manufacturers can harness the potential of AI-driven operations.

Below is an edited excerpt from the podcast:

Scott Achelpohl: Rajeev, can you lean into some of your ideas about AI and its impact on industrial network infrastructure?

Rajeev Shah: So, before I go into the wireless edge part, it's probably important first to just start with what we believe and what we are observing that AI is an evolution from what has been a 10- to 15-[year], maybe even decades-long, journey for digital transformation. So, it’s not linearly [that] we were doing connected assets, and we were learning about them and now [AI] is the next step. This is a quantum leap.

And what AI in the last … 18 months has started doing in these places is truly significantly different and changes the way these companies do business. And I want to take a couple of examples.

You can imagine being a person who is actually working on the manufacturing floor or in a refinery and you've probably had some limited ability for a decade-plus to show up at your facility in the middle of all that steel and all of that machinery with a tablet, and when you are doing some sort of maintenance on a machine, you have the ability to probably pull up some sort of a document that is like a user manual that describes how it works and so on.

That's probably not new, but just imagine trying to do that when you are on an elevated crane with all your hands just completely taken up by all the tools, possibly to even pull out a tablet and then read through what probably might be a PDF a few 100 pages long, find the relevant information, and then use it right then and there. So, while digitalization is great in the real world, in these types of environments, that's not feasible.

Now imagine that world in today's world of Gen-AI, when the same operation can be conducted through a hands-free handset, which is connected through some sort of an Android iOS device that has a Gen-AI assistant running on it, where you can ask a very specific question.

That says tell me what's wrong with this machine and the Gen-AI assistant can give you specific information looking up the same potential manual being trained on the same manual and you can right away go take action. That is a quantum leap in what that exact same information in a PDF could do before and can do now.

SA: How has the role of connectivity changed in enabling this new wave of AI-driven industrial innovation?

RS: To continue from where I left off, when we started six years ago, we didn't have this lovely term, but credit to NVIDIA for defining what is now the “age of physical AI.”

I think it's very clear that the next quantum leap, especially in the industrial environments we are talking about, is going to be physical AI. To just understand this one level deeper, I think it's worth just understanding how the tool that we are probably all using in our personal life, ChatGPT, came about. That tool that has become so powerful essentially was born because they could train their foundational model or their LLMS (large language models).

Podcast: Best practices for utilizing AI agents in industrial workflows

On the data that was already digitized on the web. Now think of what we need to do to make the age of physical AI possible. The ability to train models is clearly now there.

What's missing is how do you get the data? What's the equivalent of this big World Wide Web that was there for training on language? How do you generate that for physical AI? Where are you going to get the data to train models on your manufacturing plants or your refineries?

That, by definition, means we need to dramatically accelerate digitization and connectivity of all these old assets. And that's the first problem because a majority of these places we are talking about, think big chunks of metals, steel, a lot of liquids, lots of changes all the time, where we have essentially struggled to provide connectivity in the old world.

So, if you marry these two questions that connectivity is a mandate, so that we can start training our physical AI models and it has been hard with legacy technology. That is the first and foremost thing that we have to bring in a new wave of technology to solve this problem. The second thing that is happening is non-technological.

I think what is really happening organizationally is this is forcing companies to address the organizational gap between what they have called IT and OT because you have traditionally kept them very separate and now you really need to start bringing the expertise that today sits in your IT organizations on networking and communications and all of this and really bring it to fore in the OT world as well. And I think that's where we see the role of connectivity changing to say what it does.