How Pharma Manufacturers use the Digital Twin, AI to move Faster – Podcast Transcript
As pharmaceutical and life sciences companies work to accelerate development processes, digital technologies offer a path to greater efficiency, adaptability and speed. The Digital Twin and Industrial AI are expected to become particularly impactful.
In this episode of the Future Ready Podcast from Siemens, listen to experts on the pharmaceutical and life sciences industry dive into how digital transformation speeds development, the modeling of molecules with the Digital Twin and how AI is already influencing drug discovery.
Listen to the podcast or read a transcript of the conversation below!
00:00:13 Conor Peick: Hello and welcome to the Future Ready Podcast from Siemens. My name is Conor Peick and I am a marketing writer at Siemens, as well as one of the few hosts that you will find on the Future Ready Podcast feed. You can join me for conversations focusing on digital twin technology and adoption, as well as deep dive conversations with experts from various key industry verticals. Today, we’ll be diving back into a discussion on the progress of digitalization in the pharmaceutical and life sciences industry with our expert guests. Joining me again are Maria Grahm, who’s the Global Vice President of Life Sciences at Siemens, and Andy Whytock, Head of Market Strategy and Thought Leadership, also from the Life Sciences Group at Siemens. In part 2, I asked Maria and Andy how digitalization changes drug discovery and development, how we can go about constructing the digital twin of a molecule, and how AI is already influencing the discovery of new compounds. Thanks again for joining us, and I do hope that you enjoyed the discussion.
00:01:02 Conor Peick You know, this is kind of the point where we get to move on and talk about, okay, what can they do now to sort of confront those challenges and overcome them? So Maria, as we talk about digitalization in life sciences or digital transformation for that matter, I’m wondering what actually changes across that discovery, development, and production life cycle that you outlined for us earlier. What are customers asking Siemens to help them achieve with digitalization? And how does a connected digital enterprise in the pharmaceutical industry improve their speed, their resilience, and their decision-making across the life cycle?
00:01:45 Maria Grahm: Yeah, thank you. Now, as I described earlier, with all these different stages and in all these different stages, digitalization can make a big difference of this challenge with time. And here we feel this more and more customers coming and asking us on what can we actually do here. And I would say in each and every step, there’s something we can do. Last year, we had a great opportunity at Hanover Fair to really showcase this with our end-to-end solution already starting in the lab on what are the different steps there where you can decrease the time and become more competitive with the help of digitalization and taking it all across through the tech transfer and then also the scaling up and going all the way to manufacturing and in the end also the packaging and then distributing to the patients. But I sometimes, when my family ask me sometimes what I’m doing at work and I’m trying to explain it, and also about the challenges that the customer face, I tell my family, it’s a little bit like when you are at home and you are baking and you want to make a new bread and you experiment first with what kind of ingredients do I want in this bread and you mix and match to find maybe this is the good mix in the beginning and that would be like the discovery phase. And then of course, because you want to invite the whole family and all your friends, then you need to sort of make it bigger. You need to scale up that recipe. And for everyone who’s been baking, you know that when you use one egg in the smaller batch, you cannot just scale that up times three, because then the different ingredients works differently with each other as you scale up. So there you need then in the development phase to find the right balance between the ingredients. And then of course, when you come to the end, then to, if this was like a commercial bread making, then you need to make it in big scale and then you need to do the packaging and all of that. And it’s the same with medicine. So in all these steps, we can speed up and make more informed decisions with the help of the connected digital enterprise. Because you, coming back again to the discovery phase here, with the help of being able to collect this data and have it in a platform, we are not starting from zero. So, when we come back to Andy mentioned previously about the different insulins. So then you can base your, you want to develop a new type of insulin, but yet then you don’t have to start from zero. You can have the digital representation of that medicine already in the early phase and see what should a promising molecule look like. And then of course, like I’ve come back to it several times about the tech transfer, where the digital enterprise also can help you, that you connect. And I think you said previously that there’s this good collaboration, and maybe the collaboration is something where we can improve or our customers can improve, because we seek quite a siloed approach at our customer side between the different departments that maybe the researchers are very focused on what they want to achieve and then they sort of hand this over to. So maybe the connection is that data backbone that will really help them to also improve that collaboration in the future. And then, of course, speeding up the process. And if this then continues all the way into production, then you will be somewhat quicker in scaling up and that historical data can help you to do that, to see what has been successful in the past and do the analysis of that. And while you are in production, I would also say that if we take it even further, that what the digital enterprise can help you with, if you also can get production data in form of what is the environment in which I’m producing in, how can I identify deviations faster with the help of AI models, et cetera? That is also something that is really going across the life cycle with this connected digital enterprise will really make a difference as we step by step introduce this for our life sciences customers.
00:06:38 Conor Peick: Yeah, in some of the other topics we’ve covered, you do see this idea of data come up again and again, right? The companies are really eager to find ways to better connect their data from, either internally from their various departments and teams, and then even externally a little bit with their key suppliers, their other partners. So when you mentioned that the tech transfer, I think is the term, that has so far happened a lot through even manual, PDF type document transfers. And so now maybe moving to a more digital approach, we’re able to collect that data and move it in a more structured format in a way that’s easier for the next person to pick up and start running with. And then you also have your historical data for that matter. And so bringing that into a more digital format makes it easier to reuse for future development, so that’s yeah, really interesting how you see those parallels.
00:07:44 Andy Whytock: Yeah, sorry to interrupt you, but I mean it wasn’t just PDFs, it’s physically bits of paper that are moving around, right? These data silos really, really exist. And the data is not even available to be read to be able to do that, at least in a PDF. Sometimes you can actually read that and use it. But so there’s a whole sea change around being able to do that. To tech transfer, right? It’s about the, we’ve shown how we want to develop, like Maria was explaining, how we want to develop, or this is the ideal recipe, ideal process. Now how do I scale that up and produce that at scale? This takes two to three years, 20 to 25 months process to be able to do that. And usually involves about 25 to 30 people, different cross-functional experts. Now, of course, not necessarily full time. There are different tech transfers going on at any one time. But the time and the efforts that are needed there significantly add to the delay of bringing a medicine and to the cost of bringing that medicine to market. So we talk, it’s not as simple as just let’s make the data available electronically. There’s also cultural change in that too, in terms of what’s needed and how we do that. But having clean and clear access to, sorry, clear access to clean data will certainly accelerate that. And that’s what we see as a real game changer for the pharmaceutical industry is being able to use that data all the way along the value chain, creating that data fabric.
00:09:12 Maria Grahm: And if I may build on that, Andy, because that is a super important topic for, if we talk about digital transformation, technology is just one part of it. New ways of working, cultural changes for the people that are in this industry as well. For leaders of pharma companies, I would say this is also something that really needs to be considered when if you want to have a successful transformation into this in this environment.
00:09:46 Conor Peick: Certainly, yeah. That cultural aspect is a huge, huge part of the transformation. I think you guys are absolutely right to call out that it’s not just technology, but speaking about more technology, just to get back into our comfort zone. So, Maria, we do see that the digital twin, this idea of the digital twin is really core to digital transformation across a lot of industries. I’m wondering how can digital representations of molecules, processes, and production systems come together over time in the pharmaceutical life cycle to create a true end-to-end digital twin of a pharmaceutical product?
00:10:29 Maria Grahm: I mean, it’s interesting to see how this topic around digital twin has developed over the last years, because I’ve been in many customer meetings where we start by what do we actually mean when we talk about the digital twin, because we need to understand where is the customer also on their journey when they talk about digital twin, what do they mean? And there have been very many different interpretations of this. But I would like to say that it is this vision that we see with a fully, like the fully comprehensive digital twin of all aspects of the product, of the production, it is a journey. And we start, like you mentioned, of having the representations first, you know, part of it is in the lab, part of it is in the manufacturing space. And I would say, I expect that this is something that will grow. So we start with, how should I say, in limited scopes of these digital twins also, because it’s also a learning curve. But these molecules, eventually, these will be connected. And that’s, again, why this digital backbone is so important, that we then have the platform for these different data sources and these different representations through the value chain where they can all be connected. So I would say it’s a journey, but it’s starting. And more and more, I would say I see the change happening now in the last few years of really coming from just talking about the digital twin to really applying it and giving those good examples. We had one customer who’s also gave the great example where they have now with the help of going to continuous manufacturing on a certain product had gone from something that normally took a batch, took three days, had gone down to 20 minutes. So this is also, in a limited space. But I think it’s also for, if you are a pharma company, if you haven’t already started, I think you need to get into it and get familiar with it and decide where in my plant can I use this technology and apply it and then scale it up. Because it’s always like that. You start with it in a smaller way and then you scale it up as you are more and more successful. But here, we at Siemens would be very happy to join you on that journey and give advice because we have had these experiences ourselves. And even though we are not a pharma company, we have factories as well and manufacturing and great experiences.
00:13:25 Andy Whytock: I think Maria’s point about the definition is really, really key. I remember 7 or 8 years ago being at an ISP conference and giving a presentation around what is a digital twin. And I was really giving the very Siemens sort of centric sort of Siemens a global view of the digital twins. And it was really sort of informing people. They didn’t get really what it was about. And now it’s everywhere. And maybe we call it slightly different things, but it’s this idea of using the data and simulating and creating the real, you know, the digital world with the real world, all this type of stuff that we’ve been talking, that we as Siemens have been talking about for many, many years. And also, as Maria points out, eating ourselves. Yeah, we’ve been delivering those digital twins of factories and of equipment and stuff like that for our own factories. And I think that we talk about the digital twin of the process, the equipment of the plants. We’re now talking to our customers about the digital twin of the patient himself or herself. Yeah, I don’t know what gender a digital twin of a patient would be. Would it be a he or a she? I don’t know. But the point being that, you know, that there is, and we do do that, by the way. Yeah, you will simulating part, not maybe a whole patient, a whole body, but the heart or parts of a body to see how a drug will do that. And that’s what simulation’s all around. And a digital twin is a nice branding or a nice word that’s being used to explain, not coming from Siemens, I mean, but is a technical world that’s covering what we’ve been doing for many, many years before the term digital twin was even adopted here in the industry, simulation, modeling. This is something that the pharma company has been leading the leading global industry on in some ways, especially in R&D. How do we tie it all together? That’s the key. How do you get real value from it? And especially when you come out of that simulation part in R&D, but into actual plant design, plant building, and so on.
00:15:21 Conor Peick: It’s an absolutely fascinating thing to think about, a digital twin of my own heart or my own circulatory system and introducing a a digital representation of a therapy and then being able to watch how biologic systems react to that stimulus, I guess you could say.
00:15:42 Andy Whytock: Sorry, I was going to say, I mean, the real values that the digital twin as a whole has been seen is, I mean, in our discrete industries, you know, in an aeroplane wing or something like this, something that’s got very specific and fixed physical properties, you know. But now we’re seeing an adoption over the last five to 10 years of bringing that to these, to processes, to biologics, to different things as well, where there’s an element of uncertainty, of behavior, because of the way things are growing and so on and so forth. And that’s where we see our pharma companies, especially in the process part, of course, developing that. And you’re right, the digital twin of your body is perhaps a scary thing.
00:16:23 Maria Grahm: But I also like to think about what you simulate as well. And for instance, how you administer a medicine can also be simulated. Take the example of an inhaler. This is something that is also a great way of having a digital representation of how will the body receive the medicine that has been produced. So it’s not just by the ingredients themselves. It’s also how it’s then being received by the body. So there’s several different ways of look at this, the digital representation aspect of pharmaceuticals.
00:17:02 Conor Peick: Yeah, that’s such a great point too. I can think about all the different ways that we administer therapies and medicines. And yeah, you’re absolutely right. There is many different ways we can leverage the digital twin to sort of, test, validate, and try and work out as many issues up front and do all these things that we tend to think about when we, think about use cases for the digital twin. But so moving forward, I kind of think of the digital twin maybe as the first step into the future, right? It’s maybe laying the groundwork or building a strong foundation to continue to move forward. And perhaps you could say the next step is implementing AI. We know it’s a big topic in lots of industries right now, but Andy, I’m wondering how, or I suppose what I should say is that we know AI is already influencing drug discovery and development. It’s just something that I think many people will have seen in the news here and there, but it seems that with AI, generally, organizations sometimes struggle to move from pilot programs to scaling that up to larger applications. So I’m curious what you’ve seen that distinguishes companies that can successfully scale AI across their R&D. And then what sort of long-term impacts that this can have on development, the cost of drugs, which we’ve mentioned is pretty extreme. And of course, the speed of innovation.
00:18:30 Andy Whytock: Yeah, I think it’s the next step of what we’ve been talking about up till now. And AI is, again, it can be such a broad thing when we talk about AI in pharma, in the same way as digital twins in pharma and life sciences. I think that for many years, again, the pharmaceutical industry had been using AI for research and development, for specific analytics and stuff like this. But now AI is a much broader topic. And when you’re talking about, or when we’re talking about pilots and using AI in a more comprehensive way, then this is where pharmaceutical companies are really looking to do that. Ten years ago, there was a real move to have chief digitalization officers, right, in the pharmaceutical industry. They’re still there in some way, and maybe they’re now AI officers, I don’t know. But what I’m saying is that I think that the key is that AI has to be treated as a core capability. It’s not just an experimental thing and a nice thing to have. And I think this filters down to everybody, even to us as individuals, about how we use AI. You don’t just use it just to check your emails or to make your emails better, use it effectively. And that is the big challenge for the industry. We come back to the regulations, we come back to the how are decisions being made, how is it the, how can, how much can I trust the output from this AI engine? And there, of course, we get into the idea of the data, the workflows, the things that are behind it. So just using AI, of course, is not the thing. It’s actually about having that data foundation. We call it the data fabric sometimes as well, but having, we talked about it earlier, clean data, having data on which you can base your decisions. AI, of course, it shortens development time, whether that’s of a product or process or of the building of a plant. Of course, it reduces the number of failed experiments or failed drugs. So we can see the potential for that because it’s just making things so much more efficient. However, it does have to be treated carefully. And I don’t think anybody’s in the doubt about that. It’s the how do we make sure we treat it carefully to get those benefits. One of the key things that we’re seeing coming from regulators and the HDA is a, especially when we talk about manufacturing, is something called the human in the loop. And what that means is that the human is still responsible for any decisions. If there is a problem about a decision that’s been made by AI in your plant about the parameters that you made for your production machine, it’s still the human that makes that final decision. So AI is today being used to inform decisions that people can make and they have to be confident in the data and how they’re doing that. The move towards agentic AI, your decision-making bots and so on and so forth, this is still somewhere away because we have to have that 100% confidence in the decision-making. And this is where we get into super interesting debates because actually a human is not 100% infallible either. Right. We put those processes in place to make sure that the decision is as close to perfect as it possibly can be. And then we put the safeguards in to double check it. Once you lose control and track of the data that’s being used to inform that decision through AI, becomes so much more challenging. Yeah. So this is the big question for the industry moving forward. And I’m really focusing in on the manufacturing part here, I think. Yeah. Because this is where the control, because testing of the drugs and how that’s funded, that’s then tested in people, it will still always be tested in humans to some extent for efficacy and safety. But for manufacturing, we have to be sure, like I said earlier, and I gave those reasons why, that we know that the medicine that we’re giving to a patient is 100% going to be what it’s supposed to be. Yeah. And, you know, would you like to take a drug that’s been manufactured on AI or one that’s had people behind it taking those decisions? At the moment, the FDA and the industry believes that you need to have that human in there, the human in the loop. But how do you find that equilibrium, that balance between using AI to make that faster and better and faster decision-making to be able to be even better at it? And customers are really doing this, yeah, today. And they’re not using it to take shortcuts. That’s important. They’re using it to inform their decision-making better.
00:23:07 Conor Peick: Thanks again for joining us on the Future Ready Podcast. We will have more from this discussion with Maria and Andy on the feed soon. If you enjoyed the discussion, I encourage you to subscribe to the feed, as we will have many more conversations with experts from various backgrounds, diving into the most important trends and technologies in industry today. So thanks once again for listening, and we hope that you’ll join us again soon.
Siemens Digital Industries Software helps organizations of all sizes digitally transform using software, hardware and services from the Siemens Xcelerator business platform. Siemens’ software and the comprehensive digital twin enable companies to optimize their design, engineering and manufacturing processes to turn today’s ideas into the sustainable products of the future. From chips to entire systems, from product to process, across all industries. Siemens Digital Industries Software – Accelerating transformation.
