EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
How AI is supporting pharmaceutical manufacturing and supply chain
In this episode of the EY Health Sciences and Wellness Experience podcast, host Patrick T. O’Sullivan discusses how AI is currently being used to support pharmaceutical manufacturing and supply chain.
In this episode of the EY Health Sciences and Wellness Experience podcast, host Patrick O’Sullivan welcomes Gary Hartnett, Global Platform Leader, Small Molecule, Johnson & Johnson, to dive deep into the world of pharma manufacturing and artificial intelligence (AI).
AI has a large role to play in the future of health care. It is no longer a question of “whether” but “how” AI will transform pharmaceutical manufacturing, commercialization, and ultimately the delivery of life-changing medicines or medical technologies to patients.
Pharmaceutical companies that are adopting AI and automation technologies can indeed provide numerous benefits such as improved operations, reduced labor costs, and increased productivity and efficiency. However, to fully harness these benefits, the shift must go beyond simply integrating new technologies into existing systems. A holistic approach involving organizational transformation and change management is essential.
Key takeaways:
Leveraging AI capabilities in pharma manufacturing and supply chain will actively encourage innovation in product development, providing patients with better treatments to meet their health needs.
It is often seen that the introduction of AI and automation can lead to employee anxiety around job security. However, with the correct change management strategy, these technological advancements can be turned from seemingly threatening into valuable tools that enhance employees’ capabilities and importance within the organization.
The integration of AI and automation technologies can offer multiple advantages to pharmaceutical companies, including streamlined operations, workforce cost reductions and heightened productivity and efficiency. Nevertheless, to truly unlock these benefits, companies must look beyond integrating these new technologies into existing procedures and structures. It requires a holistic transformation that addresses every aspect of the organization, from system processes to human resources and culture.
For your convenience, full text transcript of this podcast is also available.
Announcer
Welcome to the EY health sciences and wellness experience podcast series. A series dedicated to exploring the trends that are reshaping the industry. Organizations worldwide have recognized the need to put health and wellness front and center. The health care industry has taken the lead, with a real focus on how technology, innovation, and collaboration is giving the traditional health system a radical, much needed overhaul. With so much happening at speed, keeping up is almost impossible – until now. So while everything around you continues to move at pace, take some time out from the day-to-day and join us to examine and embrace the age of health experience.
Patrick O’Sullivan
Hi, I’m Patrick O’Sullivan, and I’m responsible for data analytics, emerging technology and AI in life sciences at EY. And today I’m very excited to be joined by Gary Hartnett, Vice President of Large Molecule Partnership and External Supply at Johnson & Johnson. Welcome, Gary.
Gary Hartnett
Yeah, great to be here, Patrick.
O’Sullivan
So Gary, making disease a thing of the past, that’s an incredible credo to have. And again, doing my homework prior, I noticed that you spent practically your whole career in the world of pharmaceuticals, MedTech and then back to pharma again. So, naturally, making disease a thing of a past must have a very personal mission for you as well.
Hartnett
Yeah, it sure does. If I track back to when I first wanted to be in health care, I remember when I was about 15, 16 years of age and I was sitting at my grandmother’s funeral. So back in Limerick where I was, my grandmother had passed away from cancer and she had lived with us in our family. There’s eight kids, so a big family. My grandmother lived with us and she passed away from a short illness, after a short illness with cancer. And we came back from the funeral, and we were having the wake back in the family house. And I remember when we all came back in, I just remember very vividly I was sitting down behind these lovely, beautifully flowery curtains at the time, and everybody was finding a seat in a space. And it was a bit somber just after coming back from the funeral. And then my other grandmother, which is my father’s mom, was suffering from Alzheimer’s late at the time. And she walked into the room at the time and she says, “Where’s Theresa?” Which was the grandmother that had ... “Where is she? Where’s she?” And it actually broke the ice at the time. Everybody – everyone broke into a little bit of laughter. But for me, it was that moment where I had one grandmother that had passed away from oncology who was very close to me, the other who was suffering from a late-stage Alzheimer’s and a couple of years later passed away from that that said to me I’d always loved science, I want to work in a field where I can make a difference to people’s lives. And since then, that’s fuelled my passion to work in health care and make a difference. So whenever I think about, when I’m getting up in the morning and thinking about coming into work – and not every morning you jump out of bed to come into work – but most mornings when I get up out of bed or I turn on the laptop, depending on what part of the world I’m in, I’m driven by that passion and that fuel to make a difference in people’s lives because I know what it means to have that impact.
O’Sullivan
Surely, throughout your career, you must have seen manufacturing change and adapt. When you look back at, all the way back to the likes of Henry T. Ford, of mass production lines all the way to pharmaceutical processing today with personalized medicines, large-molecule therapies, the new technologies – most have just revolutionized your industry. Any insights in that space, Gary?
Hartnett
In terms of manufacturing, a very simple way of thinking about how manufacturing has evolved is that “when you were in manufacturing, you always ...” If you think about how do you measure how you’re doing in performance in manufacturing? So for many years, you would measure your performance metrics such as, OK, how am I doing in safety? Have I had any accidents or do I have any near-misses in my manufacturing cell or business unit? What’s my quality like? Am I doing... Do I have any deviations? What’s my supply doing? And then, what’s my cost? Do I have any waste? And you would have a big production whiteboard, we call it a production control board. And you write all of the safety, quality, delivery costs, you write them all down and then you have these measured and you write it down every day. And at intervals, whether it’s an hourly basis, whether it’s a three-hour basis or whether on a shift basis, then you perform, record how you’re doing. And then you roll that up to a weekly basis and then you roll that up to a monthly basis and you roll it from your cell up to your business unit, up to your plant, and then from your plant to multiple plants. And all of that was done manually. And then you take the information, you might put it into an Excel spreadsheet, or into a database, Access database, and then you’re pulling all of these things, and you do a transcription and transcription and transcription and transcription. Now we just have all digital screens, all of that information is collated automatically from the database in the system. And we’re always looking at ways of improving data, right? And even now, we have a big project where we’re doing that to make sure it’s consistent between all of our internal sites and all of our external partners. So we’re consistently measuring the performance. And why do you do that? Because you want to understand how you’re performing. So if something is going wrong, you want to understand how is it going wrong and you want to challenge yourself to improve. And then you’re looking at ways of continuously improving, and you’re looking at ways of sharing practices about why are you doing well in one area and maybe not in another area. But historically, it was always very difficult to understand OK, what was that whiteboard I filled in 12 months ago or a year ago? Or the whiteboard that was in the manufacturing plant over in the US vs. the one in Ireland. But nowadays you have all of that data at your fingertips. You can use AI to trend it, to look for insights and then to take action at your fingertips. So a small example of all this – I won’t even get into the technology and the manufacturing processes, etc. – you’ll find that things have become much more automated.
O’Sullivan
That’s absolutely incredible, Gary. Jumping back to that example that you just gave. So that’s been very much an advance in the physical and as a technology that’s really progressed that, and as you said, increased quality, reduced deviations and bring the best product to the market for the patient, which is incredible. But then you took all of that, having spent so long in internal manufacturing, and now have pivoted over to external manufacturing. And you mentioned AI. So how’s AI playing a role in the world of external manufacturing, or is it? Will it? Maybe some insight would be wonderful to hear on that, Gary.
Hartnett
Yeah, well, I think if you think about AI in itself, one thing that’s certain about AI is that it has a massive role in the future of health care. So it’s not a question of whether AI is going to help in terms of development, manufacturing, commercialization, delivery of life-changing medicines or medical technologies to patients. It’s a question of how it’s going to do it. So lots of organizations like our own are really looking at how you can use AI in your manufacturing, your supply chains, your value chains, and thinking about ways about how you can adopt them and how you can integrate them. I think the key thing you have to ask yourself, and I’ll help answer the question in this way, is that “what are your value drivers?” So that’s the most important thing when you think about AI. You don’t think about AI, you think about what your value driver is. So are you trying to enhance your customer experience? Are you trying to optimize your supply chain to be more reliable, resilient? To be more agile, responsive? To be more competitive? Are you trying to enhance the capabilities of your people? And then you think to yourself, OK, I want to do that. So how do I use AI to help me enable that in an accelerated way, in a different way than I’ve done before? So a lot of organizations that I’ve seen have maybe fell in love with the technology, but not fell in love with the value driver, about the by – the “why.” And we have had elements of that over time. We do lots of tests-and-learns and you need to do a little bit of that as well to try and mature your capability and your confidence with it. So when I think about external manufacturing, we call it “partnership and external supply” for a reason. It’s about partnership. So it’s like any partnership you have. We have partnerships in our own personal lives, right? And within a partnership that you have, you need to have a win-win in your partnerships. Like any relationship, if you’re taking all the time in any one relationship, it isn’t going to be a relationship which is going to be healthy and sustainable. So it needs to be a win-win relationship. We choose the partners we work with very carefully to make sure those partners fit with our values and the direction where we want to go in. Some of the stuff that we’ve done is looking at giving increased visibility to our demand, getting better access and deeper access into how they’re performing and their supply of their components and materials. So we can better anticipate and we can better solve for any fluctuations that we might have in terms of customer demand, and they can better respond to it as well. And by having that increased visibility and transparency, you may do that through planning systems, etc., using AI in terms of being able to predict. If you take that and you view ... The first thing you have to do is you have to tap into each other and be able to share data and information in a coherent way. And the second thing is you say, “OK, I’ve got that data. How am I going to get some insights from it?” And then you do stuff such as use that data to use predictive AI to tell you, “OK, you’re going to have a problem in two months’ time. We need to make a course correction.” Or what can we do to solve for this? “I have a challenge in this node in my network. You know what? Let’s use this other node I have over here to compensate for that.” So you’re building in some resiliency in that one. But the key point of starting is you’ve got to make sure that you’re using ... You’ve got to share data and information with each other and then use the AI once you have that common layer of data to drive value and insights.
O’Sullivan
It’s really, really interesting. And I really love the part around “AI is like a relationship.” I might steal that one, actually. But essentially, Gary, working with your partners as you describe them, there must have been a tremendous change management journey to go from the traditional way of operating, hey, the external manufacturer to make the batch, you don’t need to think about it until it comes out the other side, all the way to integrating AI into your external manufacturers and partners and having those insights. So, was change management ever a difficulty there or an obstacle, or was it just a very seamless thumbs-up and let’s hit the road?
Hartnett
Of course, smooth and perfect.
O’Sullivan
I wouldn’t expect anything different.
Hartnett
No, I think with anything it’s a continuous change management journey. When you’re trying to do something different, even in our internal nodes, it’s a change management journey. And I think implementing technology is one thing, and that’s probably the easiest thing. But the adoption of that technology and the sustainability of that technology or that AI is difficult. And the most important thing in that change management journey is the “why.” So at the end of the day, we should be using AI to help us be better, to be more efficient, to be more productive, to drive higher reliability and better-quality products to our patients. And it’s about understanding the step in your process that you’re using AI in, parts of your supply chain that you’re using it in, that people understand that. Because if we just say, “Hey, I want you to, instead of doing ‘A,’ which you’ve been doing for the last 10 or 15 years, I need you to do ‘Y,’ unless you tell people, “Why do ‘Y’?” Because the ‘Y’ actually could also need work as well. Because you have to do work to ... When you actually get into the world, as you know, of managing data, data is a beast. So being able to manage that data and look after that data is a whole ecosystem in itself. So you have to put effort, energy, money, time into that. So being able to use that data in a smart way, then it takes time, takes energy, takes effort. But you’ve also got to tell people what their role is in it, why it’s going to help their role or why it’s going to be good for the patient at the end of the day. And it’s a continuous journey, that change management. But when you think about partnership, I think what you’re finding, certainly in industry and in health care, is there’s become a greater acceptance for the fact that we can’t be successful alone. You can’t be successful at everything that you do. So using partners to help you be better and helping be more successful and evolving your maturity about how you think about that partnership so it’s a win-win makes a whole lot of sense, because you want your partner to be successful as much as you want to be successful in the business. And successful means that they can contribute to making a difference for patients, but also they can make a buck as well and be successful from an economic perspective as well. I think when you have that healthy relationship, it’s good. It’s good for patients and it’s good for business.
O’Sullivan
I suppose then, Gary, on the back of that, talent must be a huge requirement. So talent must be progressing massively in that space, too, with Johnson & Johnson. Have you found that, as well, that everyone needs to have some understanding of AI, everyone needs to be a data citizen?
Hartnett
Yeah, I think that the most important thing, actually, is your senior leaders need to really understand it. And so we’ve started lots of programs in J&J where you do what I would call AI 1.0, and then there’s 2.0. It’s like modules going to college. And we’ve done that at different levels. So for example, we’ve used AI to help us, so for example, for our operators and folks making the product and scientists, we’ve used the likes of chatbots to help them. So chatbot itself is an AI, to ask questions – “OK, well what do we mean by predictive? What do we mean by this, etc.?” – the chatbot helps them. So using AI is pretty cool to educate all AI. But we’re on a journey there to try and mature our capability. I think the other part we have a responsibility for, I think in general, is to make sure that we can also re-skill and upskill. Re-skilling has always been a thing where we have looked at taking people from different industries such as automotive, etc., into health care and upskilling them on maybe the technologies, etc. But now we have to upskill everybody on AI or re-skill people. I think the other most super-important thing is that we’re really, really clear as an industry, and particularly within health care, about the talent that we need in the future. Working collectively with different countries, different academia and governments, etc., to make sure that they really understand the needs in the future, whether that’s the whole way through from primary to secondary to tertiary. How we educate people. How we use AI to educate new folks coming through because they’re very different, right, in terms of how they take information in, etc. If you have young kids like I’ve had coming through, they’re very different about how they look at information. Educating them the whole way through so that they’re ready to come into the workforce in how we want them to think about AI. And I think the other piece of it is it’s always evolving. Even for me, keeping up with it is tough. It’s continuously evolving, evolving. You think about generative AI. If you look at ChatGPT and you go, “wow” – anybody who looked at that for the first time goes, “Wow, I don’t have to think about a presentation anymore, I can just press the button.” Not as good and bad at that, but that’s another day’s discussion. You’ve got to have a lifelong-learning mentality when it comes to AI. What I try to do is – OK, what we try to do in Johnson & Johnson is set some time aside for you to spend hours and days of learning, investing in yourself. And using that with a digital focus has been really our mission over the last couple of years.
O’Sullivan
It’s really powerful, Gary, and more organizations will take inspiration from that. You mentioned generative AI, and I don’t think we can have a discussion here today on AI without mentioning generative AI. And naturally, GenAI popped its head around 2017 and it’s taken maybe until about 2022 to 2023 to really build momentum and gain excitement. But are you guys looking into GenAI into areas of how we can impact health care, impact patient compliance, new manufacturing capabilities, or is it something still ...?
Hartnett
Yeah, I think it’s something bubbling away. I think the one thing to say is it’s here to stay. I think it’s evolving really quickly. It scares me a little bit when I think about it in general, because when you saw something like ChatGPT, which everybody talks about, and I know there’s lots of applications of it, you kind of say, “Wow, it’s so impactful.” But I think about, it’s not too farfetched to think about – if you think about the future, where you’re going to have a lot of using AI in your manufacturing processes. So, yes, we use AI at the moment at different levels of maturity in a lot more now than we did 10 years ago, five years ago. But in three, five, 10 years’ time, you’re thinking to yourself, “Can these manufacturing processes really run much more efficiently and effectively together? Can they control all of the stuff that we manually control and intervene today? Can they get better? Can they predict things a bit better?” Yeah. I think the technology is there. I think in our industry, which is a really, highly regulated industry, it’s about making sure we can do that in a very compliant and careful way, and when I think about manufacturing and supply chain. And I think it’s coming. We are looking at ways of doing it, but we’re also building our muscle. When you think about AI, as you know well, and generative AI, you need to have the foundational stuff really, really strong. So building your foundation really, really strong so that you’re really good, democratized data, you build that up so you have a really, really strong infrastructure and ecosystem for your data. And then you’re using AI in areas that are going to drive value for you and you’re demonstrating that. It’s a muscle that has to be built over time and a capability and a competency. And then you get to a certain level and then it’s about using maybe the generative AI to bring you on the next step. I think going from zero to generative AI in your manufacturing cost and supply chain is a big bridge, a big jump. But we are using that, and we are looking at it from a patient perspective as well, because if you think about patient adherence in terms of medications and stuff like that, there’s lots of ways you can use that help to make sure patients have a better experience, ultimately have better adherence to the medications they’re taking, and then ultimately have better outcomes.
O’Sullivan
That’s absolutely incredible, Gary. And I think that, definitely, as you said, five, 10 years is going to tell a lot for both the pharmaceutical and health care industries and the adoption of AI. Someone recently was talking to me about AI and said, “Do you think our jobs are going to be taken by it?” And I said, “AI won’t take your job, but people using AI will.” So it’s definitely ... Definitely watch that space. So Gary, unfortunately, I think we’re nearly at time today. The time has flown, but any closing remarks, anything you’d like to touch off in the space of AI or health care, broader or external supply?
Hartnett
Yeah. I think that one last thing I would say about AI is that there are things we have to manage with AI, as well. I think that there are a lot of super, super-potential value enablers through AI and we can see them and we’re on a journey and we’re going to continue to invest in that journey over time. You need to make sure that you take care of data privacy. I think it’s really, really important because you may be looking at personal or financial data, etc., and that’s quite sensitive, there’s a lot of regulations and policies. Cybersecurity is a risk when you have a lot of data, etc., so making sure you have systems to protect that. And then costs can be prohibitive as well as an entry to smaller firms and companies, because it takes quite a bit of capital investment to make sure you have the right infrastructure and ecosystem in place that we can do it. You’ve got to be conscious of the things that you need to take care of, as well, to enable AI. But that being said, AI for me is a game changer. And it’s only going to continue to, I think, ultimately to make us better in terms of being able to ... When I say better, it’s about being able to supply high-quality products to our patients. It’s going to help us make more innovative products faster. If you think about it now, even in the development of our products from an R&D perspective, we use AI for screening of new molecular entities, so we are able to do it a lot faster. We’re able to pick those products that make a difference quicker, those new molecules, the new compounds that we need that are going to treat diseases of the future. So on our mission to create a world where disease is a thing of the past, we’re helping using AI to help accelerate everything from the patient piece of it, so that they have more awareness and visibility and more access, but also to the development of those right through the supply chain, so that we can actually make more higher-quality, more reliable products, get it to the patients in the way that they need it, but also get more innovative products to them as well so that we can continue to make a difference in their lives, which is, as I said earlier, is what fuels me and fuels everybody in Johnson & Johnson.
O’Sullivan
Gary, that’s been incredible. AI is a love story, as Gary highlighted, and AI supporting making disease a thing of the past. Gary Hartnett, Vice President of Large Molecule Partnership and External Supply at Johnson & Johnson. Gary, thank you so much for the time today.
