In this episode of Great Question: A Manufacturing Podcast, Scott Achelpohl is joined by Nick Hasselbeck and Matt Martin of E Tech Group to discuss the challenges legacy OT systems create for digital transformation and AI adoption. The conversation explores why siloed data, limited network visibility, and outdated equipment remain major obstacles for manufacturers. They also examine practical strategies for connecting older assets, evaluating the value of stranded data, and balancing equipment replacement with lower-cost modernization approaches. The episode offers a grounded look at how manufacturers can build a stronger data foundation for AI and other advanced operational initiatives.

Below is an excerpt from the podcast:

Scott Achelpohl: With our last podcast, we focused on positive vibes and OT and IT convergence success stories. With this one, we're going to be a little bit more real about the challenges that legacy (translation: older) OT systems present for digital transformation. What's without a doubt the most hyped and popular digital project right now? That would be AI adoption.

We're just as guilty as anyone, we've been writing all the time about the promise and, beyond the exuberance, real use cases at some companies for agentic AI, meaning virtual workers plugged into plant operations besides flesh-and-blood technicians and line operators. We've been doing a lot of myth-busting about AI adoption this summer and really all of this year.

Thank you very much, gentlemen, for joining us. And let's jump right into Q&A. I've got some questions prepared, and we'll see if they yield any more follow-ups. Here's my first one: Is the blind spot we discussed in our opener primarily caused by the fact that legacy OT equipment is not networked sufficiently and thus provides no visibility or very little or is unable to be networked? Nick, I'll let you start.

Nick Hasselbeck: I think that's part of it, the physical aspect. I think another part of it, though, is the history of the industry and how the join-up between IT stacks have matured that contributes to the blind spot.

And, so, when I think about the industry over the last couple of decades, the thing that jumps out to me is that there have been several waves of transformation.

Going way back, you have ethernet or the trend toward converged plantwide ethernet. It's kind of the first big transformation. And you have virtualization coming into play closer to the plant floor.

And then more recently, you have the security posture evolution and threat landscape changes. And what strikes me is that the threat landscape changes that have been common in corporate IT are increasingly a mainstream obstacle or opportunity as you get closer to the plant floor.

So, even if it's physically able to be networked, that's the first hurdle you have to step beyond. But the second one is given the threat landscape or given the environment that the plant is in, if it can, is it wise to connect it to the plant? First, is it possible? Then is it wise? And if it's not, how do you make it so or what needs to change?

Matthew Martin: I think it's a mix of both, across the sites I've been at. Recently, that has been mostly a life sciences focus as food and beverage was more early in my career. So, most of my answers will kind of be life sciences focused. And there is an interesting mix of both modern and legacy equipment in these facilities.

The big current pushes, in addition to the AI side, are integrations to MES, LIMS, like those kind of enterprise or business planning software. And then for the AI enablement, it's like how do we get data into a data lake so we can kind of let loose our data analysts, business analysts to kind of get value for the business?

So, as Nick alluded to, there's kind of two categories. There's can we do it at all or is it physically impossible? My definition of “physically impossible” is that there's literally no ports, no connectivity, no wiring of any kind that you can tap into off of the device or the equipment.

And in that case, you kind of have two options. You either replace the equipment with something that gives you the same function or better that can do the same thing. Or a bit of a more hybrid approach is like you could augment that existing machine or hardware with some sensors or capability.

So, one thing that comes to mind, it's top of mind for a lot of people, is preventive maintenance. If you have an older piece of equipment and you want to get it into some preventive maintenance program, you could throw thermal cameras out there, vibration sensors, that can more easily plug into a more modern data architecture and kind of revive that system or at least bring it in and get the data you need off of it.

And then yeah, the second piece, I'll just echo what Nick said, where if it can be networked, you then need to negotiate like OT versus IT, make sure everybody's understanding of the situation and that there's an agreement there.

SA: Another question that came to light: If OT equipment is able to be networked at all and prepared for AI piloting and later implementation, what's the best way to go about that networking project?

NH: Oftentimes, what we see is that the first step is to do an assessment. Some environments already have a robust asset inventory, already have the right mix of stakeholders from IT and OT and the business involved. But that's not necessarily always the case.

And it's common to go into a site where there's a range of stakeholders and a mix of known and unknown elements in a given facility. And, so, the first step is really to get the right mix of stakeholders involved.

Again, typically IT and OT but in particular a representative of the business. And so most often that looks like a decision-maker that has KPIs that they're responsible for and or has knowledge of a data set or knows that there's high-protein data that we that you can that may be high value that can we connect to a large language model or connect to an AI work process that is currently offline and or not connected and getting an organization to come in and do that kind of assessment, to step through that asset inventory, help identify either assets that need to be networked or risks that need to be remediated.

And ultimately, a partner who can articulate a risk profile and make recommendations that speak to the business requirements and that address either the risks that the organization is concerned about or address the opportunities for attacking that AI “pilotable” data.

MM: There's two pieces to these kinds of assessments. So, you have like the physical infrastructure, which Nick kind of touched on, get the stakeholders in, make sure you understand like what assets are there, how they're going to connect, like what the backbone of your communication stack is going to be from an IT perspective.

From an OT perspective, I think the more common thing is you need to engage stakeholders and vendors in what are the options for some of the data access patterns and how to actually get data off of the equipment, assuming that infrastructure is in place? So, some obvious kind of common patterns are OPC UA server, it's probably one of the most common ones for OEM equipment.

It could be an MQTT client or an MQTT-compatible device that just goes in there and starts publishing data and subscribing. Then you're kind of more classical. You've got proprietary drivers for your big PLC brands. And then you kind of have some more of the edge cases of it could have a local device database, it may or may not have an API to access it.

It could be generating flat files that you then need to push over the network or scrape off the device. So, each of those factors kind of dictates how you can realistically integrate it into your more modern data ecosystem.

About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.

Listen to another episode and subscribe on your favorite podcast app