The Life Sciences industry is undergoing transformative changes as personalized medicines and unique therapies are developed and brought to the marketplace. Companies are driving digital transformation efforts to move toward autonomous operations, meeting the needs of people who require these innovative therapies.
In this podcast episode, I’m joined by Emerson’s David Zhang to discuss these digitalization efforts and the important role that real-time scheduling plays in bringing these therapies to market in a timely manner.
Give the podcast a listen and visit the DeltaV Real-Time Scheduling section on Emerson.com for more information on advancing towards autonomous operations.
Transcript
Jim: Hi, everyone. I’m Jim Cahill, and this is another “Emerson Automation Experts” podcast. Today I’m joined by David Zhang to discuss DeltaV Real-Time Scheduling, or DeltaV RTS for short, and how it helps manufacturers drive improvements in their operations. DeltaV RTS was recently honored with Plant Engineering Magazine’s 2025 Product of the Year Gold Award in the Automated Processes category. Welcome to the podcast, David.
David: Thanks. Thanks for having me here, Jim.
Jim: Well, I’m looking forward to our discussion. I guess to open things up, can you share some of your background with our listeners?
David: So I guess how far should I go back? Maybe 20 years ago, we started this program out of UC Berkeley. I was a Ph.D. student at the time in industrial engineering, and one of our desires was where could we apply industrial engineering to other industries? For those of you who don’t know, industrial engineering is all around simulation, optimization, throughput maximization, and it’s typically used for places such as airlines, semiconductors, logistics, etc. And what we saw was, Life Sciences, was actually a place where people are still using a lot of manual processes. And we thought, hey, is there something that could apply there? And the company that we founded called Bio-G was born. Fast forward to about six years ago, we were acquired by Emerson, and now we are one of Emerson’s scheduling simulation platforms.
Jim: Well, that’s great. We’re so glad when you joined the family there. So let’s get into the meat of things here. And let me ask you about BioPhorum’s Digital Plant Maturity Model. I know given all the changes among manufacturers and the Life Sciences, that companies have a huge focus on assessing and advancing the digital maturity of their operations to better meet the challenges posed by all these changes. What are some of these changes and the challenges they present?
David: And I think maybe we can start talking about the Digital Plant Maturity Model. For those of your listeners who might not know what it is, it’s really like a guideline to help a company assess where they are in terms of their digitalization journey. And you can really think about it from level one all the way up to level five, one being more of like a pre-digital plant, where there’s a lot of paper-based processes, low-level automation, PLC-based controls. And as you move up, you’ll hit things such as the level three connected plant, which means that companies usually have a lot of software already implemented, vertical integration. They’ll have things such as an ERP, an MES. And then at the furthest along of the digital maturity model is what’s called the adaptive plant. And the adaptive plant really is meant to be a lot more aspirational. It’s what’s called lights-out manufacturing where there’s very little, if any, human input when it comes to manufacturing products. It’s fully end-to-end value chain integration.
And I think a similar example is if you think about the automotive industry, that’s what’s called fully automated self-driving. So right now, you could say companies like Tesla are sort of there. You have the autopilot, I think what they call it. That’s really budding towards level five. And for some example, like the Waymo, the RoboTaxi, that truly is fully automated. And that’s the reason why you could say customers want to get to that point is there’s a lot of benefits getting there. When you take humans out of the loop, you have increased consistency, you have lower safety issues, increased quality, lower costs to make the product. And all of these things are applicable to what we see in the Life Sciences industry today, which is trying to maximize the amount of product they can produce for some of these life-saving therapies, as well as different types of modalities. There’s things such as personalized medicines or CAR T, which we can talk a little bit about later, but they present different unique challenges that having a fully adaptive automated facility can really help with.
Jim: Yeah, I think that the self-driving cars was a really good analogy there to aspirationally where we’re trying to get to in our manufacturing processes. So what role can real-time scheduling play in advancing digital maturity towards BioPhorum’s level five, this adaptive plant?
David: Right. And if you really think about what adaptive plant means, it’s fully automated. And I think it’s a good exercise to think about how you actually get there. And if we take a step back where the reality of everything happens is the shop floor. Those are the actual processes that are required to make the product. And then as you go up from the actual process itself, there is automation. And this is essentially a software that controls the equipment, directs people what to do, right, in order to make the product.
And that’s where we come into play is we take it to the next level. So this is what’s called like bottom up, where we come from the automation layer upwards. And if you think about automation as it knows everything it’s done in the past, for example, the historical information, and it knows what’s happening today, and we take the baton and pick up from today all the way into the future. And you can see if you continue on that path, all we need next is a feedback loop that takes the information that the schedule is saying and push that back into automation layer, and boom, you have the fully automated lights-out facility.
The key difference here with DeltaV Real-Time Scheduling is the fact that we’re bottom-up. And the majority of scheduling platforms out there, they’re really supply chain tools. They’re planning tools, meaning they’re coming top down. And unfortunately, you can never get to lights out if you do the top-down approach because it’s a clear disconnect between what the schedule thinks the plant should do and what the plant is actually doing.
Jim: I guess specifically from DeltaV RTS, when we’re talking about DeltaV RTS, how does it help improve production throughput?
David: And I think that’s a little bit of a different approach that we take vis-a-vis other platforms is RTS actually has a second component to it called DeltaV discrete event simulation. And really, that simulation is what feeds the schedule. So the schedule is essentially an output of the simulation. And what we do is we make a model of the facility. And this model really needs to contain all the information that is used to run the facility.
For example, you want to capture the assets, all the equipment. Obviously, you want to capture things such as people, utilities, anything that can be rate limiting, you want to capture. And most importantly, you want to get an idea of some of the more esoteric or hidden information. What I’m talking about our things such as tribal knowledge, if someone’s been running a plant for 20 years, they have a lot of information on what to do and what not to do. Right? And the model needs to capture that information. And similarly, any kind of automation conflicts, if there are certain things in automation code that would prevent a facility from running as expected for whatever design reason, you want to capture all that information as well.
And the importance of doing this is when you think about increasing production throughput, it’s all about the bottleneck. You want to identify where the bottlenecks are, resolve those bottlenecks, and then go on to the next bottleneck to try to resolve. And if you don’t know…if you don’t capture that information in your model, there’s no way it can be identified as an issue. So I think part one is making a very high-fidelity model of the facility.
Now, to get value out of that, you need to have the integrations. And that’s a key portion here is it’s really this type of garbage in, garbage out. If you don’t know what the facility is doing today, then you don’t have a lot of confidence in trusting the outputs of the model for all future work or future predictions. And that’s really where we come in is we have a direct connection in automation. And so, therefore, all of the results that are being produced, the user or the customers have a high confidence that those results are very accurate. And that’s really in a nutshell is making a model, wanting it a debottlenecking exercise, and then evaluating the results with confidence.
Jim: Yeah, I can see having that data and being able to use that against what the model is saying, you know, to really get it and hone in and get that throughput going on there. Can you share an example of how a manufacturer has driven improved production levels?
David: Sure. I can think of two examples and they’re very stark differences, which using this type of modeling approach first, I think really can help. The first is some of these blockbuster drugs. We have customers that have some of the best-selling drugs in the world or some of the GLP-1 weight loss drugs. And the common theme behind this is, A, these guys are capacity constrained. Any type of product they make, they can sell and they want to implement a system like RTS to maximize their uptime. If you don’t have this type of system, let’s say you’re still using a manual-based schedule, then it’s going to cause issues,right?
Sometimes people can’t get the facility back online fast enough. People need to figure out, hey, there’s production delay, you throw a lot of people at it. Well, that all results in less time for these customers to make these blockbuster drugs. So we’re aiming for, as much as possible, 100% uptime. And RTS is a way for customers to achieve that without having an army of schedulers, essentially, or an army of people trying to figure out the way to optimally run the plan. Right. So that’s a real value add for us.
And then the second portion is really these personalized medicines. So with personalized medicines, it’s similar to having…you know, as the namesake suggests, you have a lot of patients coming in. They’re all going to be different. And how do you process them in the right order so that you can get the maximum amount of throughput? And that’s a stark difference than, you could say, traditional manufacturing, where if you’re trying to service a larger patient population, you just make bigger and bigger tanks. Right? With personalized medicine, that kind of approach is not their work. It’s each patient is unique.
And I think the analogy I like to use for this, and I’m a big analogy guy in case if you haven’t realized yet, is it’s around like airports. And if you have airports, they really have the same type of problem. It’s about throughput where you have all the planes coming in. They’re going to have different information in terms of when they’re going to land, if there’s delays, if they need to do maintenance or whatnot. And if you try to use a manual schedule, it’s possible to run the airport in such a way. But it’s obviously not going to be optimal and you might have a bunch of disappointed customers. A lot of people waiting on the tarmac or waiting in the terminal for quite some time. And what do they do? They have an algorithm that automatically feeds all this information in and then it’ll reassign when the planes are going to leave. It’s going to reassign the gate, right, it’ll be a new estimated time for departure. And that’s exactly what RTS does is it takes all this information in and then it tries to get the patients out as quickly as possible.
And I think that’s also another unique thing about this is, at least here at Emerson, we obviously want to help our customers improve their bottom line. But the unique thing in Life Sciences, we have an opportunity to actually impact patients. And especially with these personalized medicines, these customers are very sick. And there’s a terminology in the industry called vein to vein where they want to get treatment from vein when they extract the patient cells, manufacture the cells into a therapy and get it back to the patient within some certain amount of time. And that’s what RTS does is it helps our customers ensure that they can meet their vein-to-vein times with a high degree of confidence. That’s something that’s really not achievable without having a fully-automated system like Real-Time Scheduling.
Jim: Yeah, that’s a really good example there of just whatever you can do to shrink that time and what it means for the patients in there in the personalized medicine case. So I guess from a scheduling perspective, how does DeltaV RTS address unexpected issues that may occur during the course of a batch production run?
David: I would say the easy way to think about this is RTS, because it’s an engine and it’s an algorithm, it will automatically recalculate the best course of action in case of a delay. And it’s very similar to a GPS. If you’re trying to get from point A to point B, what does the GPS do? It takes in the current information or position and if a road is out, if you want to take a rest and break, make a wrong turn, it’ll recalculate a course for you and it’ll give you a new ETA. And that’s something that we found has a lot of benefit in terms of adoption because these issues are going to occur in any type of manufacturing facility, in fact, any type of scenario at all. Right? There’s always going to be some unexpected events. And instead of spending the time trying to adjust course, what we’ve seen is by the time customers adjust the course, the batches are already moved on. Right? So things are already happened. So it’s imperative that you have this type of real-time calculation, much like a GPS that helps you do all this stuff automatically so that the site can get back to business as usual instead of being in this firefighting mode.
Jim: You are very good with analogies. That’s a great one to help picture, you know, how you course correct just based on what’s happening in the process or with the equipment, everything else there. Can you share an example of how this has been used to improve operational performance?
David: Yeah, we have a customer which has a relatively straightforward process. It’s what’s called like a serial process. And when they manufacture this therapeutic, it’s basically you do A, B, C, D, E, right? Of course, there’s supporting operations around that, but the overall flow is relatively straightforward. And you might want to ask, why do you need an advanced scheduling program for such a thing? So previously they were using basically Excel. And you’ll see that a lot of customers today are using Excel. And if you have a serial process like this, the ease of Excel is that you can do whatever you want. And it’s easy to move the blocks around. And when there was a delay in the process, let’s say there was a two-hour delay, and all you need to do in Excel is move the operation out by two hours. Right. So really easy to use. But what happens if the inverse is true? And now something happens two hours earlier?
Well, Excel doesn’t know that. And that’s exactly with our integrations, RTS does know that. And so it’ll scoot the schedule in by two hours. And what that means overall is now you can produce more product or more batches in the same amount of time. And so that helps us increase our operational performance and increase the throughput that these customers are getting. And again, in turn, what that means is if I can get a 10% or 20% increase in throughput, that’s 10% to 20% patients who now get access to this therapeutic that might not have had it in the past. So again, that’s something that we always like to think that at the end of the day, much like our customers, it’s not a site that’s getting the benefit. It’s ultimately the patient. And that really helps us drive for more efficient work. And it really helps us create an ethos of why we’re doing things here at Emerson.
Jim: Yeah. Really knowing that you’re not just driving greater throughput, you’re helping more people in the world. So that’s a really good way to think about that. Now, I have to ask, what does winning the gold-level award in the automated processes category mean to you and the entire team that built up and supports DeltaV RTS?
David: Yeah. I mean, it’s a huge honor, for sure. And I think I speak for myself as well as the whole team here at Emerson that it’s great. It’s a great recognition for us. Only took 20 years or so to get here, but we know that especially in Life Sciences, things move very slowly. And I think people are sort of used to that. But it’s great. It’s definitely a validation of the things that we have been doing here that is really making a difference. And being able to get this gold-level award, which was a huge surprise to us actually, we didn’t even know that we were in the running for such a thing, is a very pleasant surprise and a very pleasant thing for us that I think already has definitely improved the spirit here at Emerson. And it gives us confidence that whatever we’re going to do in the future is also something that is great to be the gold standard for.
Jim: Yeah. It must mean, if it came as a surprise to you all, that your customers were talking it up and getting the word out about it, which is a great thing, which is what you want. And I guess on this 20-year journey, now technology is so rapidly advancing. So what are some near-term product enhancements and how will they help these manufacturers?
David: Yeah. I think the sort of default response to get from many software vendors is AI. And we’re no exception. I think the way we think about AI a little bit differently is we want to figure out where AI can be applied to a place that’s going to have impact, right? And no AI just for the sake of AI. And I can really think of two areas, at least in the short term, where we can see AI having an impact. The first is really around sort of like an AI co-pilot, which can help people, engineers, schedulers that are building these models, create models at a much quicker pace. What I mean by that is when you think about how things are currently done, if you’re trying to create a model, you might read some documentation, you might consult one of your peers on how to do this. And a lot of this now can be replaced by AI, right? Let’s face it, nobody wants to read this huge manual on how to do things. They just want to have AI scrape that information and give you the result.
And more importantly, it can help you write the code or write the model configurations on how to do this. If you’re thinking about, hey, how do I model this type of cell therapy or how do I model this type of, you know, media prep or buffer prep, it’ll spit out the configuration requirements or the code that you need to do so. And I think that encourages a way of having best practices where we can have the AI read all of the best practices, the standards that we recommend to customers, and then we can export that information, right, to basically anybody who wants to use it so they’re not trying to reinvent the wheel. And I think that really helped increase the quality of these models as well as reduce the time to deploy and the cost of maintain all these models. So that’s really the first part.
I think the second part is really having AI change the way that we use or interact with user interfaces. Because a lot of…I mean, at the end of the day, what people want are outcomes. And the traditional way of getting those outcomes is you need to interact with the UI. If I had an RTS open, right, I might use a schedule, I might drag and drop stuff around. But why am I doing that? It’s because I want to see what the impact is, right? I want the output of that information. And what we have building right now is RTS AI bot that essentially does that for you, where you can take a lot of information and instead of having a human do it, you have an AI do it. So for example, let’s say Jim, you’re my shift supervisor, and you say, “Hey, tomorrow I’m going to be sick. I’m not feeling so well.” And what I can do is I can ask the AI to do that. Say, “Hey, Jim’s going to be out tomorrow. How does that impact my production for the next shift?” And it’ll recalculate things for you and give you an optimal outcome to say, okay, if you’re down one person, this is what it looks like, right? And that’s really the agentic part of this where we want to have the interface for customers be the AI. Instead of having to learn the software and then you click the wrong button and it gives you the wrong answer, let’s just have AI do all of that. So I think those are the areas where we feel like AI can make a big difference.
Jim: Yeah, that really makes sense. You want the outcome and the traditional UI, you’re trying to click to get to that, what you’re trying to do. And yeah, I can see how AI can really streamline that whole thing, as well as the first example that you gave. So yeah, those are some tremendous uses looking for it. I guess maybe looking a little further out, how do you see DeltaV RTS evolving to meet future demands in plan automation and scheduling and enabling advances in digital maturity towards the adaptive plant?
David: Yeah, I would say there’s some really interesting use cases right now. And we’re definitely in the right time and the right place where there’s good reason to go towards the adaptive plant. The scenario I talked about earlier, the personalized medicine, the cell therapy, that’s again a scale-out process, which basically means that as you’re trying to service a larger patient population, you also need a commensurate manufacturing organization that can support that. So if you’re going from 100 patients per year to 10,000 patients per year, you’re going to need a proportionate amount of manufacturing personnel to achieve that.
And of course, that is fraught with issues. One, it drives up the cost, but you might not even be able to find these types of people. Personalized medicine is a relatively new sub-industry within Life Sciences, and there’s not a lot of qualified people who can actually work on the manufacturing floor. So for us, that’s great. And what this has culminated into is basically having a fully-automated cell therapy facility where it is lights out. And the justification for that is to reduce the cost of goods sold. And I’m happy to say that we’re involved in some projects that we’re helping our customers achieve that. And that really is the bleeding edge of the bleeding edge. So I think that’s a very cool and exciting time to be in.
And I think that the second portion of this is it’s something that Emerson as a whole is trying to move towards. Some of your listeners might have heard about Project Beyond for our enterprise operations platform. And really, what that is is to create a data fabric for the entire enterprise data from the OT layer to extract all that information. So you have all of your MES systems, your PLMs [product lifecycle management], your schedulers. Right now, that information, even though it is there, it’s not necessarily unified. And what we want to do is extract all the information out.
So now the flow of data is easily accessible by any type of platform that wants to use it, both Emerson platforms and non-Emerson platforms. That ensures that this type of access is scalable, there is security all built in, and then it creates a perfect environment for AI to access all this data and send commands to the right system so that you truly get this adaptive plan. And I think that’s one thing that has really changed the game here is that AI now exists and it’s sort of like household name now that we’re trying to set up our customers to be able to fully visualize or realize that vision using our enterprise operations platform.
Jim: Well, David, I want to thank you so much for sharing your insights. For our listeners, make sure to check out the DeltaV Real-Time Scheduling section on emerson.com and the DeltaV for Life Sciences section on YouTube. Thanks for joining us today, David.
David: Thanks. It was a blast.
-End of transcript-
