Natan: Hello and welcome to augmented ops. We have an exciting episode today. We have with us Jim Fox from AstraZeneca. He recently moved up to the US to take the position of senior Vice President of America's supply Chain operations. So we're very excited to have you on the show. Welcome, Jim. How are you? Very good, thank you.

Great to see you. So today we're gonna take a deeper tour into how complex biopharma environments lend themself to digital transformation, which is like a huge topic and we've been working on it. AstraZeneca has been working on it, but give a bit of background. How did you end up in this crazy industry?

Jim: Almost by luck actually. I did a PhD in chemical engineering and I was sponsored by a different sector. And I went into that sector and I was there for two, three years. But it was at a time where pharma was really growing and the sector I was in was starting to decline. What was the first drug or uh, technology you were working on in pharma?

It would've probably been something like, uh, ranitidine, which is, uh, an antacid when you look back, 

Natan: what is it, three decades of a career. Yeah. And look at all the therapies and drugs you helped create. You must look back and say like, wow, we helped a lot of people. 

Jim: Yeah. The mission of most pharmaceutical companies is to serve the patient.

Yeah. And to make lives better. 

Natan: Yeah. A lot of people are cynical about that sometime. 

Jim: Well, you know, it's a business as well. It's a business. Yeah, for sure. But if you're an innovator company, like my previous company and AstraZeneca, you know, we're looking to solve unmet medical needs. 

Natan: Mm-hmm. 

Jim: When I was in my previous company, GSK, I didn't think at the time we were gonna find a solution to HIV.

Yeah. You know, in the late eighties people were dying from this. Absolutely. You know, I was involved in some of the drugs that basically took that off the table. So you know, it still, the hairs on the back of my neck go up on that. Yeah. Because you know, we solved something that was a death sentence.

Natan: That's not true at all. Where is AstraZeneca? Tell folks who don't know a lot about AstraZeneca, a little bit about the company and where you all focused. 

Jim: So I've only been in the company four years. So you're a noob? I'm a newbie to this one. Yeah. But you know, there was a reason I joined the company.

Mm-hmm. Both GSK and AstraZeneca, both UK headquartered companies, actually quite similar in nature. Mm-hmm. But I was really sort of interested to see how the, the share price. In AstraZeneca was sort of almost going exponential and really curious as to why that was growing wildly. Why, why, why, why. Yeah. In AstraZeneca, the ambition to really grow the business and the pace at which we do stuff and the collaboration that we do was really quite interesting for me.

I. 

Natan: So from your perspective, I think when people say digital transformation, you know, the concept is well understood, but at the same time is is abstract. Mm-hmm. And I started telling anyone who would listen to me that we should actually stop using that term. Yeah. Interesting. Why I. Because where we are now in 2025, that's probably not as been true 20 years ago.

And Windows 95 is rolling high, or you know, 30 years ago. So really digital transformation is happening. But in 2025, I asked myself two questions with respect to digital transformation. Mm-hmm. One, what is a non-digital transformation? In other words, digital is like obvious. You cannot do anything in this world that is not digital.

Mm-hmm. And the second thing, when is the transformation over? And I argue that it's actually not over ever, because as soon as you're done with one piece, then there's another. So at the risk of coining another term, and I don't expect you or the audience or anybody to like, uh, follow that. Mm-hmm. I started talking about continuous transformation, meaning that you drop the digital.

And you make it non terminal. So you get continuous transformation, which obviously has like a dialogue with kaizen and continuous improvement and all that. Yeah. But I want your perspective, like how have you seen in pharma mm-hmm. This concept of digital transformation change over the decades. 

Jim: Yeah.

That's really interesting and I totally relate to what you're saying. 'cause if you put 20 people in the room and ask them what their definition of digital is, you'll get 20 different answers. At least. At least. Yeah. So it's, it's a little bit unhelpful, I think. And it's all things to all people. When I first started, we didn't really talk about digital.

We talked about automation. Automation, and that was more sort of the, the industrial automation, the OT site. So back in the day it was more around, especially with API plants, it was how do you make the plant safe? 

Natan: Yeah. 

Jim: In a drug product plant, it was more around how do you get the discreet control of a particular unit process.

You know, in pharma it, it takes several steps to get to an end product. But 30 years ago, we didn't connect those steps. 

Natan: So it wasn't like sending Johnny with a pH meter. It's like the pH meter is actually attached to the equipment. Yeah, that's right. And sending reads over SCADA all the time. And that was like, oh, we're, we've automated that.

Jim: Yeah. But the interesting thing is we were collecting a lot of data from those sensors, but we weren't really doing much with it. And in the first instance, that's I think how we started to define digital in a different way, which was how do you make better use of data? Yeah, and you start with visualizing it and then you start looking at trends, then you start making some decisions.

Not necessarily in real time, but you know, that's where the human process engineer and you training people exactly what to what, what does it mean, what does 

Natan: it mean to all this data? 

Jim: Yeah. And that was really at the process data level. 

Natan: Yeah. And Astra, you just recently moved to the us Yep. So congratulations on the new role.

What, what is your new role? 

Jim: So I'm heading up, uh, what's called America's Supply Operations. Mm-hmm. So it's both North and South America. We have seven manufacturing sites and it's really to look after the supply of products, both coming into the country, going outta the country, and what we make within the country and the US is our biggest gray market.

Natan: But part of that, you were in Sweden and the risk of embarrassing you, I know you guys won for Sweden Global Lighthouse Network Award for one of the sites there, correct? Mm-hmm. It was modeled for operational excellence. What made that a lighthouse site? What merited the recognition I got? What did you guys get right there?

Jim: So, if I go back a little bit, the first opportunity I had for Lighthouse accreditation was actually back in GSK in 2018. We were one of the first farmer sites to be accredited, and that's a little bit back to what I was saying with the digital world. A lot of what we put forward was around making sense of data, looking at trends, making data available to all sorts of people in the organization.

But year on year, the bar from the Worldwide Economic Forum was getting harder and harder to join that, that club, if you like. Yeah. So by the time we started to look at this in Sweden, visualization of data was a minimum entry level. Activity. It's like table stakes. Absolutely. Just a means to do business.

So what we had to do was think differently about how you convert data using processed twins, for example, and mathematical models to start predicting some of the performance attributes from physical properties of drug substance. 

Natan: Whenever I hear those stories about, you know, using predictive type of algorithms in hard industries like pharmaceuticals, you know, highly regulated.

Yeah. Many ways to get it wrong. Yeah. This issue of trust in the technology comes to my mind. Mm-hmm. Not only externally. So like you adopting a piece of technology needs to trust the technology that the people who actually end up using it need to trust it. How did you work on this trust? How did you.

Help the organization take in all this tech and actually change how they work and trust that this is actually better, faster, you know, all the attributes you can attach to this. 

Jim: So really what I'm talking about is, is how you accelerate development to a commercial product. So a lot of these algorithms were developed within our development organization, 

Natan: just clinical, 

Jim: started at clinical supply, but then moved into sort of the tech transfer process.

Right. Interesting. Yeah, so the piece that stimulated this really was we had an accelerated opportunity with a particular product, so we wanted to get it filed as quickly as possible and launched. And we didn't have much drug substance to play with. And that was typically how, you know, in pharma, you'd do a series of design of experiments at small scale and then you would tech transfer it into a commercial facility.

Natan: That's always fun. Easy. Yeah. 

Jim: No 

Natan: couple of weeks. 

Jim: No couple months. 

Natan: Couple Three years. Three years. Roughly. 

Jim: Roughly. Roughly. Yeah. 

Natan: Huge bottleneck in the huge 

Jim: bottleneck. Yeah. And it's simpler to model liquids. Mm-hmm. At scale it's harder to model powders. Powders, yeah. Yeah. So what what was really interesting, are 

Natan: you talking as a chemist now?

Uh, more of a chemical engineer? You're talking the behavior of the substance? Yeah, the physical properties. Really. Physical properties, yeah. How'd you get them to flow? Yeah. Why is that important? Help us non chemist understand why it's so critical. 

Jim: So there are two parts to making a product, right? That you have the active ingredient, which is the bit that does, does the job, does the magic.

Yep. Then you've gotta convert it into a, a. Dosage form. Yep. That typically it's a tablet or a syringe or something like that, but you have all of the other ingredients that go with it in order to make it effective. So there's more than just the drug substance in a tablet. So when you're talking about powders, you've actually got to.

Mix the powders together, and some of those powders are high density powders, some of'em low density, so it's a little bit like trying to mix coffee granules with baking powder or something like that. So you have very different properties that have to come together in a mix, homogeneous way to get 

Natan: the perfect, the perfect mix for your coffee.

Yeah, 

Jim: exactly. And life doesn't always work simply like that. So you've gotta try and have these sort of process steps that get the mixing and then the flow of powders between unit operations. And that's not as easy as it sounds. That's why we tend to go into, you know, granulated products. But the trend going forward is not granulated products.

'cause the more steps you put in, the more energy intensive they become and less economic, the 

Natan: more control that you need to. 

Jim: So what you want to do is really have powders that flow. By first intent. Yeah. Yeah. So what we did in this particular, and back to your previous question, is we started to mechanistically model how the physical properties of the different powders interact, and therefore how you can get the right combination to flow through a a unit process.

Natan: So this might be a bit provocative, but when you look at the industry, how do you see the speed in which, and the back to the pace here we are back to the pace that all this digital technology and the change that ushers into an organization. Are you happy with the pace? 

Jim: That's, that is a tricky and provocative question.

Yes. Because it's a highly regulated industry. Right. And you might say, well, so is the nuclear industry, so is the aerospace industry, da, da da da da. So 

Natan: is transportation. And so is, you know, airport security and 

Jim: Yep. We live in a regulated world. We, we do. We. But I would argue when you are designing molecules that have biology changing effects in human beings, well, there's always safety, for example, with the aerospace industry.

But when we talk about clinical safety in human beings, that's of paramount importance. Mm-hmm. So, you know, there are three things really that in our industry we look for. Safety, is it tolerable? Is it efficacious? And can we make it available? Mm-hmm. So if you launch a product, you gotta make sure that you don't short the market because people have been waiting for this disease changing experience.

So you need, you need scale. You need scale for sure. So it's always a little bit of a tension between the regulators that oversee good manufacturing practice. Mm-hmm. Versus the speed at which, you know, if you're in production, you want to do things more efficiently with time. 

Natan: But let's, for the discussion here.

Assume we completely satisfy regulation. Yeah. How do you get, I guess, digital at scale? One of the interesting things I've learned in our journey with Astra was that there was like an edict or a mandate or some sort of a, I don't know, like thou shelt, that I've learned from you. A, make sure that if you're doing something lean, it needs to be digital.

And if you're doing something digital, it also needs to be lean. Yeah. Explain that decision to mandate that kind of approach. 

Jim: A little bit, like you said, digitalization, we're all sort of Corning, the lean digital piece. Yeah. Which becomes a little bit of a thing, but if you break it down, what it really means is, and this is going back in time a little bit 'cause I, I think it changes over time, but Yeah.

Really you should follow lean principles first. What you don't want to do is automate or digitize waste. That's a good one. You don't want to automate or digitize waste. So if you have a lean mindset, you simplify first. Mm-hmm. Then you standardize, which is a business process. And then you automate or digitize.

What I don't think you want to be doing is digitizing bad process. Bad pro. Exactly. So it's the opportunity to lean your process first from a business perspective where it's working, and then you can make it super fast with the digitization. 

Natan: Yeah. You know, we're seeing this phenomenon across a lot of pharma customers who are regulated, that they have literally thousands of SOP Yeah.

SOPs for all sorts, like SOP, how to Scratch your Ear. In a proper, don't underestimate that pathway way. I, I, it's a totally, like, especially in a bunny suit, you know, you need to know how to do that properly. In the tension there, it's like, on one hand those, so p they're like the Bible. 

Jim: Mm-hmm. 

Natan: I'm not underestimating how critical it is.

Then it's in the QMS and it's there and it's solid. 

Jim: Yep. 

Natan: But when they want to take those so p and turn it into like a compliant work process, they're kind of like. We don't work like instruction as a PDF, that like humans don't work with like 1, 2, 3. They work in a lot more state of flow. Yeah. In reality.

Yeah. That doesn't necessarily violate all the requirements of a FDA regulated environment. Mm-hmm. And that has been a very interesting journey to see how people take to that and really transform sometimes decades of SOPs. 

Jim: So I'll give you a view and it's a personal view. Yeah. In most companies, the quality management system is actually straightforward.

At the highest level, and it makes a lot of sense because they come from, you know, international regulations and, and they're pretty clear. Yeah. What happens over time is we make the SOPs a little bit more complicated, especially at the local level. Why is that? Because you, you have this tension of purity.

Versus, you know, a recent inspection. And that could be from an external agency or that could be from our own internal self-inspection. Mm-hmm. So typically what happens is if there are any sort of findings within inspection, the first go-to thing is, well, we'll update the SOP. We will update the SP and then then you get layering.

You get layering, and then over time they just become really hard to follow. In other words, compliance. So in words, they're 

Natan: saying like QMS is a tough thing to maintain. Well over time, especially when you're adding more products, you're adding more tools. You've got to watch out for the 

Jim: creep. The creep and the complexity that comes into it.

Do, do you 

Natan: think quality leaders in the industry recognize that? 

Jim: Oh yeah. Absolutely. What 

Natan: are they doing? What do you see people do to 

Jim: approach that? Well, I, I'll say in our company, uh, it's a really proactive call to organization. Mm-hmm. They really are starting to embrace digital. Mm-hmm. So we moved a lot of our SOPs into one platform.

Mm-hmm. Which is the start of then simplifying and improving. So, because you had multiple QMS reality or something? Yeah, multiple repositories of where we stored these documents. So one of the first things to do is put it in a single repository. Mm-hmm. Globally, you mean? Globally? Globally. In one repository.

Then you can start to do the SOP 2.0. Which is how do we then start to rationalize those? How do we start to take out the complexity? How do we make them more fit for purpose? Mm-hmm. And then of course, once you've got them in a good state, then you can start doing sort of data mining or information mining.

As you say, there are thousands of SOPs. You imagine, you know, an operator. Or a first line leader or someone 

Natan: holding it in their brain. No way. You 

Jim: can't do that. Yeah. And typically in the past, we've trained people by read and understand the SOP goes back into a a repository. Tell me 

Natan: that. You read that, 

Jim: tell me that.

Read. Yeah. Tell me that you really read that. Yeah. And of course people did. I. But then the retention of the exact procedures, et cetera. Yeah. You know, you need to have a refresher quite frequently. 

Natan: Yeah. There's a lot of research on that. 

Jim: Yeah. But now of course, with trying to disseminate some of the regulations from work instructions, which are slightly different, and having those practices or work instructions in an agile way close to where operators are doing the work.

Yeah. And reinforce 

Natan: the SOP without being like a digital version of that one-to-one. Correct. And that brings me to the second big term that I think is interesting to unpack because we've been hearing it across the industry. A lot of people are now focused on autonomy or creating autonomous manufacturing environments as opposed to we're just gonna adopt automation that usually, you know, the mind goes to whatever robots and this and that and you know, lights out and we heard all those terms.

Mm-hmm. What is an autonomous. Factory. What is that for you? What is that for a company like AstraZeneca, 

Jim: uh, that's really interesting. I, I actually prefer the word automated. Automated. Mm-hmm. That's the first one. Well, it comes from the literal translation in Japanese of, I think it's Judoka, which means automation with the human touch.

Natan: I think we got our, uh, episode title right here. Yeah. 

Jim: So, you know, this isn't new. This is what Toyota were doing. Yeah. Many years ago. Yeah. Personally, I think in pharma, certain operations lend themselves potentially to lights out, but in the main, when we look at drug product facilities, that is not the way to go.

I still think you need some human intervention as you start turning powders into product or liquids into final product. 

Natan: Well, the FDA will tell you you need the humans, at least 

Jim: in the foreseeable future, right? Yes. That I think is the difference. So for me, autonomous in a factory environment. Really, we work from warehouse to warehouse, so the incoming goods from the raw materials, raw materials, consumables, you name it, all the way to, we've now got a packed product ready to go out to the patient.

But there's one thing in the pharma is you can't just do the physical flow. It's got to be in lock step with the release process from quality. So you can be as really efficient as you like with the conversion of raw material materials into product raw. Yeah. But if you don't have the quality approval at the same pace.

You've got leg. So for me, autonomous is all about how you create smooth physical flow and smooth information flow such that when the pack goes back into the warehouse, it is released and ready to go. So the sounds simple. Yes. Easy to say. Not always easy to do because you've then, I mean, what we call a value stream is you've then got to get all the components of the conversion steps in the right sequence or cycle time.

So there's a, there's a whole combination of how you turn those powders into product, and there's a little bit of waiting occasionally, so you wanna minimize the waiting between process steps. But of course, at the moment. You know, a lot of it is offline testing to prove that the quality is right. So you've got to coordinate the QC as part of the flow within the process.

Natan: Yeah. This idea of orchestration of the state of production and set value stream. Yeah. Like when do you do the inline quality? Yep. When do you do release? How do you do deviation and exception handling right there on the floor? Like to me, an autonomous operational environment, let's just, 'cause it could be a lab, it could be a warehouse.

Like you said, it's necessarily just the factory. Correct. And the cleaner it needs to talk to the humans there and tell them, Hey humans, and now we need you. Mm. And the humans need to be able to talk back to the operation. And I think like just a lot of architectures in the past 20, 30 years actually made sure the humans never talked to the technology that actually like very few people from the purpose of tight controls.

And, you know, 

Jim: e bit like black box, wasn't it? 

Natan: A lot of black boxes and a lot of frankly, like old architectures. Yeah. That many systems, their principles were put together in a world that doesn't have iPhone and everyone walks around with like the ability to do pretty sophisticated things and have a trust in the tech.

Yeah. I'm not speaking about trust and Yeah. How that applies to how we augment a pharmaceutical process. That is what we're working on. We're making good progress, but it's certainly not done. But you know, I think that's the paradigm shift that is happening. And you know, some of the stuff we've been working on together, I think exemplify that.

Jim: Well, a really good example, and this was part of our Lighthouse accreditation, I. Is the platform by which we convert powders into product, in this case, coated tablets. Yep. So where the formulation lends itself to this, you can collapse a number of process steps onto one machine. Mm-hmm. So, you know, we managed to reduce the manufacturing lead time from what would've been typically two weeks to 20 minutes.

That's game changing. 

Natan: That is 

Jim: right. How 

Natan: did you do that? 

Jim: So it's continuous technology. Mm-hmm. And you're controlling every step of how the powders are flowing into the process. But don't forget, you have to have the exquisite particle flow properties in advance to make it flow in a good way. So there's a precursor to just running the equipment and then the models to allow you to set up the equipment in a good way.

So it becomes much more than just the metalwork and the platform. It's the digital ecosystem that surrounds it. Now, the QC piece I was talking about was, well, if you can make coated tablets in really high quality in 20 minutes, you kind of gotta make sure then that the QC testing doesn't become the massive bottleneck then to progression, right?

So there. You know, you're moving from what would be a sort of an offline testing approach to, as a minimum. At line and moving more to sort of online, we're using process analytical technologies. Now. That's something that I was involved with in the early two thousands with FDA, with the PAT for guidance for industry with PAT, and that was cutting edge at the time, but we couldn't really get it to stick at the time because the digital systems that surrounded it weren't as sophisticated as they are now.

So now I think more than ever it's become a reality. 

Natan: Yeah, well we could talk about it for hours. I've been telling a lot of architects recently, it's like, you know how lines are blurring. Yeah. Between the various kind of tools that do stuff and, and sort of put the emphasis on orchestration, you know, with all my tool of bias.

I think that's part of what we're helping and providing. But Jim, we hear about the revolution that biologics are actually doing. Mm-hmm. You know, as opposed to traditional chemistry with powders as you mentioned, like how do you see that changing operations? What's the impact there? 

Jim: So we shouldn't confuse sort of large molecule and small molecule with its different conversion process into the the final product.

Some of them are very similar. Mm-hmm. It's just the way you create the large molecule versus the small molecule. Small molecule is synthetic. Yep. You know, there's chemistry, large molecules tends to be more like biology. Yep. With cells and bits and pieces, right? But it 

Natan: means you have different production environments.

Jim: You have a very different production environment with very different cycle times or or manufacturing lead times, and it typically costs a lot more money to produce a bio drug substance. Just the nature by which it is created. I think I know where you're going, but what we are seeing is a, a world where we're creating much more sort of niche products that are spec or a market of 

Natan: one.

Jim: Correct. So if you take rare diseases, for example, that we're very much into, you know, you might have a population of less than 10 in the world that is, is waiting for our medicine, right? So you imagine the whole supply chain aspects of dealing, you know, clinical supply to start with. Sometimes these people are needing a home environment to do this.

So rather than going to the pharmacy or what you typically do, so how we change that whole supply chain is really different. But also how you manufacture it when you get to the drug product side, it's a much smaller scale. I. 

Natan: In smaller sites and different machines. And 

Jim: so if you go back to my previous site instead, Italia, you're going from a world where we were making over 10 billion tablets a year of medicines for let's say CVRM, cardiovascular, renal, metabolic.

But then when you get much more into the niche disease states, oncology, for example, and so on, you're down to the tens of millions. When you get into rare diseases, you're down into the less than 10 and named patients. 

Natan: That's a good note to end on because we started talking about scale, but people forget that you say scale, say, oh, I need more.

Yeah. But sometimes the same operator scale means like you actually need less and you need to design production systems that can support the flex up and down. 

Jim: The key is agility and flexibility. Yeah. 

Natan: Jim, thank you so much for joining. Very welcome. I said augmented ops has been great, and I'm looking forward for your next lighthouse.

I'd like to see how you guys evolve your production system, so yeah, appreciate you. Thank you so much. 

Jim: Very welcome. 

Natan: Thanks Naan. 

Announcer: Thank you for listening to the augmented Ops podcast from Tulip Interfaces. We hope you found this week's episode informative and inspiring. You can find the show on LinkedIn and YouTube or at tulip.co/podcast.

If you enjoyed this episode, please leave us a rating or review on iTunes or wherever you listen to your podcasts. Until next time.