AI in Biotech: Hype, Reality, and the Future of Drug Development

Sabine Hutchison
Welcome to Cut the Chat — Life Science Insider.

Season 2, Episode 4: AI in Biotech — Hype, Reality, and the Future of Drug Development.

In this episode of Cut the Chat, Life Science Insider, I'm interviewing Christina Busmalis. She's a leading expert at the intersection of AI and biotech, and together we'll explore the transformative potential of AI in drug development — discussing its promising applications and the challenges that still need to be overcome. Christina is going to share insights from her experience, offering practical advice for biotech companies looking to explore possible uses of AI in their processes, and looking at future trends. This episode explores the reality versus the hype of AI in biotech, providing a wide-ranging perspective on how this technology can shape the future of drug development. So let's get to it and cut the chat.

Christina Busmalis is an accomplished technical and sales executive with more than 30 years of experience, focused mainly on the healthcare and life science industry — specifically big pharma and biotech. Most recently, Christina was Chief Revenue Officer at BenevolentAI, an AI-augmented drug discovery company, and she's also a non-executive board member of the AO Foundation, a nonprofit organization dedicated to improving the care of patients with musculoskeletal injuries and pathologies. Christina has also held executive leadership positions at Google Cloud, IBM Watson Health, and IBM. She's a computer scientist by education and a consultant by trade — an industry thought leader on disruptive technologies for the life science industry, and a frequent speaker at pharma and digital-health-focused conferences across Europe, where I first met her. In my chats with Christina, I've found she just explodes with innovative ideas, and loves to create solutions for her stakeholders, focusing on digital, data, analytics, AI, and cloud transformation. Christina happily lives in Zurich, Switzerland. So welcome, Christina — it's great to have you on Cut the Chat.

Well, hello, Christina — it is so good to see you again, and I'm just thrilled that you're with me today. We met earlier this year, I think June, right at the HBA conference, and I was amazed at your ability to take a topic like AI and make it so engaging. I remember listening to people at my table, and they were saying, wow, she's doing a really good job making this palatable and understandable for everyone. So kudos to you — I don't know if I ever told you that after meeting you the first time, but everyone was very excited about your talk. So, really happy you're here, and that you've taken the time to be on Cut the Chat.

Christina Busmalis
Great, I'm happy to be here — really excited to have this conversation.

Sabine Hutchison
Yeah — and off we go, into this wild, wild world of AI and biotechnology. This is your world right now, isn't it? I think you have a long history in this — I'm really excited about this topic, at least the data and technology side of it. So we're really looking forward to hearing a bit about you — where did it all start?

Christina Busmalis
Yeah, it's interesting — I started in tech when I was quite young, eight years old. I started to program, which — giving my age away — was back in the early 80s. My father was a math teacher and a computer science teacher at high school, and every break he'd bring home a computer. We didn't have any games for it, so I just had to play with it in different ways — did some basic programming, built lines across the screen, which back then took a lot of lines of code, now it's done for you.

I just loved technology, so I studied it through college too, and moved into technology — not the typical tech path, I guess, since I actually didn't want to stare at a screen all day long, which is funny, because nowadays we're on a screen all day long. But I'd say, over my career — how did I get into the AI field? I actually did some AI programming back in the 90s, when I was still in university.

Sabine Hutchison
Right. Yeah, just see different things, don't we? Right.

Christina Busmalis
But that was in early stages, very early stages. I'd say there have been five key career pivots. One was coming to Europe in the late 90s — I came to Zurich, which should have been a two-year stay, but I'm still here, so that's a long time later. The next one was really moving away from pure consulting — IT management consulting — into more of a sales role, a sales management role, at IBM.

And really moving into the life science industry — that was probably my biggest pivot, really getting into the life science industry and seeing how much needed to be done with technology. I spent the rest of my career in the data and analytics space. Moved away from being more on the tech side, around 2017, to run the life science business for Watson Health at IBM, which was IBM's first AI-based practice.

It was great — I was responsible for Europe and Asia, doing a role I'd never done before, since I was never a sales rep — I was leading a whole team of sales leaders and sales reps, bringing AI solutions to market. I then left IBM and joined Google — so I went from old tech to new tech, and realized there's not much difference on the other side. And then I really made the jump, moving out of a pure tech company and into a biotech company — BenevolentAI —

last year, last fall. And for me, it was a really interesting move, because I moved away from being a tech supplier into the biotech-pharma industry, into an actual biotech company. But it was still a company with a tech side — they debate themselves whether they're biotech or "tech bio." Just as a two-sentence intro: they do AI drug-discovery-based solutions, taking things up through bringing—

Sabine Hutchison
Right.

Christina Busmalis
—bringing the medicines into clinical trials, and then, ideally, out-licensing them to move them forward. I took on responsibility as their chief revenue officer — dealing with out-licensing of potential pipeline assets, selling collaboration and research initiatives with pharmaceutical companies. They also had some AI-based software they'd developed themselves, which we brought to market. So I really liked the move, going from a big tech company to—

Sabine Hutchison
A small biotech company, of a couple hundred people as well. Yeah, big shift. And also this combination, right — bringing AI into clinical research. AI is just so prevalent today. I think ChatGPT kind of kicked off a lot of this, for the everyday person who doesn't focus on it. But we forget this has been going on for a long time — it's not something that just started a couple of years ago.

Christina Busmalis
That's correct — yeah, absolutely. I mean, the 1950s, if you think about it, that's way back when they started doing artificial intelligence, and the capability has just increased and gotten better and better because of data and processing. Those are the two things that have really accelerated it — especially with generative AI — having so much volume of data, and having the technology and the computing power to really accelerate that.

Sabine Hutchison
Yeah, definitely — that is so true. And what's fascinating is this whole connection between AI and biotechnology, which is something you've obviously been focused on at Benevolent. I'm curious — how is that, and how many organizations are you seeing that are actually focusing on drug development, and incorporating and running their companies based on AI?

Christina Busmalis
It's a good question, because — I haven't done the actual research, so I'll make a hypothesis, and someone can prove me wrong — I think every company, throughout the entire drug development process, has AI sitting somewhere. Whether it's machine learning, or really strong advanced or predictive analytics, AI is being used somewhere. So if you think about early drug discovery—

Sabine Hutchison
We'll have to fact-check it. True. Yep.

Christina Busmalis
—whether it's identifying novel targets — again, that's what we did at Benevolent — assessing those targets, validating them, there are AI pieces everywhere. And then if you go into clinical trials, it could be protocol design, patient recruitment, or site selection — AI is being used there too. Would I say it's brand new? No — but it's woven into pieces of existing processes today, wherever you look. Already being used in very small areas.

Sabine Hutchison
That's true — it's quite siloed, isn't it, if you look at it that way. Yeah, yeah.

Christina Busmalis
It's very siloed — I'd say siloed, or fragmented, you could say. There's definitely been a — I won't say transformation, because I don't think it's a transformation yet — there's been an implementation of AI, or machine learning, into various areas within drug discovery and drug development.

Sabine Hutchison
Yeah, and I think if you look at it — going into the clinic, and what you're saying too around clinical trials — you're right about protocol-writing support, also feasibility, looking at where, and what countries, what investigators are starting to use it to determine a lot of these points when you're running clinical trials. What about a step before that — on the development side, where are you seeing the biggest impact being made there?

Christina Busmalis
I think it's still mostly in AI for drug discovery — there are a lot of companies out there. Again, BenevolentAI was infusing AI across the discovery process — from novel target identification through to the hits, bringing the molecule to IND-ready. They accelerated the process, which typically takes five years, down to three years. If you ask where AI is being used the most, I think it's—

Sabine Hutchison
Mm-hmm.

Christina Busmalis
—in acceleration mode, at the moment. Separate from identifying novel targets — that's a second step — it's more about accelerating right now, which I think is what's coming out. But going from five years down to around three years already cuts into, reduces, two years off that ten-to-fifteen-year development cycle we see today.

But this isn't new — there are companies like Benevolent, and a lot of others out there. You've seen consolidation in the market too — Recursion recently acquired Exscientia. So there's a lot of acceleration happening, a lot of companies doing similar-but-different things — everyone's doing something slightly different, but landing in the same area. So I think that's where you've seen the most focus. Now, where I find it challenging today is that we haven't yet proven it works, because—

Sabine Hutchison
Right. Yep.

Christina Busmalis
—is that the patient? We're in a lot of phase two trials — a lot of companies, In Silico Medicine and others, have dozens of trials in phase two, dozens of molecules in phase two, but no one's yet in the market, no one's yet at that side. So until we really get them there, it's still — we don't know. Are we confident this is the way forward? Probably. But it's not yet proven that it definitely is.

Sabine Hutchison
We don't know. Yeah, that's true. And do you see also — because we're in an industry that's fairly slow in a lot of areas, and we're talking about acceleration — do you see, or have you experienced, regulatory bodies being able to handle that acceleration, and keep up with the speed you're seeing in some of the development?

Christina Busmalis
I think what's happening, still today, in my opinion, is that things haven't really changed — you still have to go through all those steps with the FDA, the regulatory authority, whether it's getting IND-ready, what you need for phase one completion, phase two, and so on. They're not blocking it, by any means — they're not saying you can't use technology to help you — but you still need to do all the paperwork, all the information required to get there. It doesn't just solve—

Sabine Hutchison
Mm-hmm. Right.

Christina Busmalis
—it doesn't just change the process — the process itself has changed, we've just infused different ways of doing things into it. I've also seen — we talked about patient recruitment — years ago, back at IBM Watson Health, we had a clinical trial matching solution. It technically worked — it would take inclusion and exclusion criteria, match them against patients, and identify the best candidates for whatever trial you wanted to run. The challenge wasn't the technology — it was getting the parties together. Having three-way contracts, with—

Sabine Hutchison
Right.

Christina Busmalis
—the provider side, the pharma side, and obviously IBM as the technology provider — that took a long time. Just to get that effort in place, you could actually be doing your recruitment in the meantime. So you have to balance the two. It wasn't that the technology was the hindrance — it was more that companies weren't ready, weren't used to it yet. That was, you know, "this is the first time we're doing this sort of thing," ten years ago.

Sabine Hutchison
We see that as well — sometimes just getting a three-way CDA signed can take months.

Christina Busmalis
And maybe — this is probably the benefit of biotech versus big pharma. Big pharma is more challenging, but can biotechs take that and move things forward faster?

Sabine Hutchison
Yeah, but it's also — if you look at it from the vendor side too — there are some pretty significant vendors out there as well who also struggle with changing processes, adapting processes, making things happen quickly. So that's a really good point too, because it's not just about the research and development side, the sponsors — the vendors are going to have to adapt to AI in the industry too. What are your experiences with that?

Christina Busmalis
So I think the CROs — if we want to say they're the pivotal piece for biotechs adapting the process. AI can be used very early on, in discovering things — that maybe isn't the CROs' domain. But once you're taking a medicine and bringing it through the clinical trial, the development process, most biotechs outsource that to CROs.

You know, digital endpoints — all those things — the big ones are doing it themselves. For small companies, it's hard, unless you partner with the big ones to get the funding. So that's the challenge they run into — it's not always about the technology.

Sabine Hutchison
Yeah, true. And it's interesting you mention digital endpoints, because we're actually working on a project with DiMe specifically about this — trying to get more biotechs involved in this space too, to start utilizing it. A lot of them already do, but also to get their voices out more around digital endpoints, and incorporating that into their clinical trials and protocols.

Christina Busmalis
Yeah, that's a good example — again, it's the technology, the sensors — we're moving a little away from AI, but it doesn't really matter, because it's all connected in some way. The sensors, there's more and more capability out there, so the technology probably won't be the blocker at some stage. It's going to be more about how you get the interdisciplinary communication and agreement that this is beneficial, and how you make sure the data is valid, accurate, and clean — those sorts of things.

Sabine Hutchison
Yeah, it's connected. Right.

Christina Busmalis
But it could change a lot for diseases where you don't have good clinical endpoints — heart failure, for example, what's your endpoint? I think there's going to be — and again, where does AI play into that? It could be in the sensors themselves, it could be in gathering all this data and getting insights from it. All of those areas are going to be linked to artificial intelligence, as well as what comes after.

Sabine Hutchison
Yeah, that's true. So true. And it's one of our previous conversations — we've had a couple of chats since we met, and we talked a bit about this: scientists not trusting — or, maybe not that they wouldn't love to trust it, but they can't completely trust AI in the drug development process. I find that really interesting too, and I'm curious to hear how you've seen it impact the process, and what you think needs to happen to change that.

Christina Busmalis
Yeah, so — trusted AI, I think that's a really interesting topic, because there's this question of, how do you trust something you haven't decided on yourself? How do you, as an individual, believe information you're given? You research it, you make sure you're confident in what you're hearing. It's the same thing with scientists. If I go back to these AI drug-discovery companies again, whether it's Benevolent or others—

Sabine Hutchison
How do you trust it? Yeah.

Christina Busmalis
—it's the same thing. Benevolent used their platform to identify potential novel targets — not for one specific disease, but based on multimodal data. Taking all this data together — publications, patient data, different types of data — you can get inferences and information that can potentially point to new targets. So the system doesn't give you one target — it gives you a prioritized list of targets, by disease area, based on that data. And then it's a scientist's job to go and validate and assess those. You can't just hand someone a list and say, here you go — that's not how it works. The process to get to that list doesn't actually take very long — it's the process of validating and assessing the list, and then going into wet labs, where the real time is spent.

From that list, a lot of it is going to get thrown away, because the potential evidence the system provides might be good, but not good enough. Or it's already being worked on by someone else, whatever the biological reason. Then, I think, it comes back to the fact that biology is complex — super complex. And I'm not a biologist, I'll never claim to be, I don't have that background.

Sabine Hutchison
Right — it's so complex, right.

Christina Busmalis
And technology is helping accelerate scientists getting there faster. I've always talked about AI — actually, back at IBM, ten years ago, we used to say AI isn't "artificial intelligence," it's "augmented intelligence." Now that's become an obvious thing to say — everyone says, we're not replacing humans, AI is there to help them do their job better and faster. And the trust aspect matters because—

—I think, especially for a scientist — if you see too much garbage, you're not going to trust it. And it's not perfect, and I don't think it ever will be. Things like AlphaFold have helped — protein modeling and folding done in advance — but you still have to validate. Scientists still have to make sure the evidence it's presenting as "good" is actually good enough.

I wish there were an easy way to say, hey, we're going to change this, and everyone's just going to trust AI — but we're not there yet, especially in today's world of generative AI, where there are hallucinations. It's manageable, but it takes humans to help manage it. That change just doesn't happen overnight.

Sabine Hutchison
Yeah. Interesting, talking about scientists — what's your experience actually hiring? Bringing in people who have that mindset too — people who think, well, we can use it, but we still need to test, we still need to validate. Have you seen a difference in the type of individuals, or maybe the people coming out of universities now — are the mindsets different?

Christina Busmalis
I don't think it's an age thing — there's always this assumption there'd be an age thing, but I don't think there is. If I look at my experience at Benevolent, we had scientists of varying ages — not super young, some young, some older, some in the middle. There wasn't really an age pattern. I think it's more about the willingness to approach things differently. We found very good people coming from other biotechs, from pharma, from university — and I think there are a lot of people out there willing to look at things a bit differently than maybe was done ten years ago. I don't have hard evidence for this, it's just a hypothesis, but I think that's probably changing more and more, with people coming out of university too.

Sabine Hutchison
Which actually fits science, doesn't it? Because it's all about curiosity — that's why I studied chemistry too, I was curious, I wanted to understand how things work.

Christina Busmalis
And I think especially the majors, the university degrees you get now — bioinformatics is becoming standard, it was quite rare before, you'd either do biology or you'd do technology, now people are coming out doing both. I think that's just going to keep accelerating. And there's still — in most of these AI drug-discovery companies, you have a tech side and a science side, and how you meld them together really matters. We were able to do that quite well at Benevolent.

Sabine Hutchison
Yeah.

Christina Busmalis
The tech people have to learn that not everything is perfect on the tech side, and the science people have to learn to start to trust it — and when it doesn't work, to communicate why it didn't work, so it can be retrofitted.

Sabine Hutchison
So how did you make that happen? Because I think, in general, it's difficult to bring two different groups, two different mindsets, together and get them to really collaborate. What was your secret sauce there?

Christina Busmalis
Yeah, I'd love to take credit for that, but I can't, really, because I came in later at the organization. I think it was leadership that really brought that together over the years — having that passion, and having the right people, a great chief scientific officer, a great CTO, those kinds of things, to bring those capabilities together. What I saw, when I joined as chief revenue officer, was more about how you glue it together to make it sellable, to drive revenue — because that was my job, bringing in the revenue side.

And that worked, because whenever we did different sales pitches and things like that, it was about getting both sides at the table — because your potential clients want to hear from the scientists as well as from the business side. It makes the story work.

Sabine Hutchison
Yeah — a diverse group at the table, again that theme of diversity comes up. So true. So — what do you see, what's in the future? If you pull out your magic ball — where are we going with this? Maybe not even super long-term, but if you look at the next five years?

Christina Busmalis
Yeah, very — I think so. Again, we talked about AI hitting the process at a pretty low level so far — a fragmentation, we haven't yet transformed drug development with AI, we're starting to transform pieces of it, different processes within it. I'd hope, over time, that we start to really — I'd love to see somebody reinvent — and that has to be aligned with the regulators and everything, it's not going to be just one company — but I'd love to see someone reinvent the whole drug discovery and drug development process. One thing I see — I don't think it's five years away, I think it's a bit longer — is digital twins. Digital twins have been around a while, in manufacturing and supply chain, simulating what you're going to produce — and you're now seeing that come into healthcare much more strongly.

Sabine Hutchison
Quite a bit, I think.

Christina Busmalis
Yeah. Imagine — today we give a therapy hoping it works. We're getting more and more, obviously with genetics, and the more we know about the patient, the more confidence we have it's going to work — but we still go through that process, especially in clinical trials, of figuring out how it will work and who it will work for. And it's a long, painful process for the individual.

When can we get to the point where we can simulate clinical trials with digital twins — up to the point where we go into patients, and I think you'll always have to go into patients, I don't think that changes — but can you go into patients already at a high level of confidence that it's going to work for that type of patient? Not for all patients, of course, but for a type of patient — with a high probability of success. I think that's when we'll have really changed something. Because maybe there won't be all those phases of clinical trials anymore. Maybe there won't be placebos.

Ideally — and being able to use more digital endpoints too, all of that comes together around that. And you're right, it's already being used — I think Johns Hopkins University is running a clinical trial creating a whole computational model of a heart, to deal with fatal arrhythmia. They have a 3D view of that individual's heart — everyone's is different — so on the computer, they simulate the operation before—

Sabine Hutchison
Right.

Christina Busmalis
—the procedure, before they go through it. And then they actually treat each individual patient differently. Again, we're not there yet, but I think that's where — can we get to a point where we develop medicines completely differently than we do today? I don't know if it's five years—

Sabine Hutchison
Yeah — maybe a little bit longer, but it is coming. I mean, look at the warp speed we're at right now. It feels like warp speed, with the technologies and what's happening in the industry. So, yeah.

Christina Busmalis
And when you think about it — even COVID helped accelerate that. There's negative aspects of COVID, of course, but there's the positive too — how it brought everyone together. But I think we've kind of taken a step back since COVID, in terms of not keeping that momentum. But let's hope we can get back to it.

Sabine Hutchison
Yeah, that's true. And to trust in ourselves, and to keep moving forward — that's so true. So — what about you, the next five or ten years? Are there areas you want to focus on more? Is there something drawing your attention?

Christina Busmalis
That's a great question — I wish I had a direct answer, since as a salesperson I should. But I'm still very passionate — we already talked about passion — I'm still very enthusiastic about the life sciences industry. I'd love to see us move faster, and I'd love to be part of that movement. I've been in the life science industry for a little over 20 years now, about 20 years —

working with AI, in some form, for about 12 years. And we're making progress, but not fast. It's funny — back at the HBA conference, there was a woman who spoke, she was amazing, and I remember, at the end of her talk, she said, "this isn't for me, because I'm going to retire soon, but for the next leaders — you have to do this." And I remember thinking, I don't want that to be me. I want to actually be part of that.

Sabine Hutchison
Right, be a part of that change, yeah.

Christina Busmalis
I do have quite a bit more time to work in this space. But I think staying focused on how technology — not just AI, because I think AI is just one piece of the technology picture — how we can continue to bring technology into healthcare and life sciences, to really help all of us be healthier as we go forward.

Sabine Hutchison
Yeah, let's say you have a few more years. Yeah, fantastic — I know you will be. It's interesting, if you think about the biotechs out there that want to start, that want to take bigger steps — what are the top three Christina-tips you'd give them, to take a bigger step into this AI space in their R&D?

Christina Busmalis
I want to be part of that journey, absolutely. If we look at the R&D process — not so much the innovation they're bringing, but the process itself — choose the CRO, if you're going down the CRO path, which is most likely, choose one that's going to be a trailblazer, or get on that bandwagon. Or partner — ultimately most of these companies partner with big pharma — find a pharma company that can help drive what you want to do, moving forward.

Sabine Hutchison
Mm. Right.

Christina Busmalis
I think that's the interesting thing out there — it's an exciting time, but you can't do it alone. If you want to move sideways, you have to find the right companies to bring into your ecosystem, and take the time to actually do that — rather than just outsourcing to whoever.

Sabine Hutchison
Yeah, so true. And knowledge-sharing — I think that's such a key thing. Sometimes we tend to hold information and great ideas close, rather than sharing them. I think that has such a huge impact too.

Christina Busmalis
Yeah, that's a great point. I think, unfortunately, the industry has driven that closed-door kind of culture. It may come down more to individuals who are interested in sharing knowledge, than to companies driving it. Because you're right — that's what COVID did — companies came together to figure something out, and looked through it together.

Sabine Hutchison
Christina, maybe we're going to have to start some roundtable discussions, or a next project, to get people talking and sharing information.

Christina Busmalis
I'd love to hear from you — from those trailblazer biotech companies — what do they want to do? How do they want to be innovative, not just in the medicine they've identified, but as a company, as—

Sabine Hutchison
Yeah, and to the greater good — you're right, to their organization, of course, that's what it's about, they have to be successful, and then ultimately it's for the patients, to bring something there. But how do we come together to blaze more trails? Yeah, I like that. All right, well, we'll have to see if that sparks — if there are any trailblazers out there who want to be part of—

Christina Busmalis
Yeah, exactly. A good trailblazer. Yeah.

Sabine Hutchison
—something, let us know. Because it's true — we hear a lot of CEOs, a lot of individuals, get together for round tables, and maybe it's about setting up that kind of safe environment where you can go — that's your place to talk about these things, and see how you can leverage each other. I think that's so important — we don't have to do this all alone. There's so much great stuff out there.

Christina Busmalis
Yeah, because the drug development process is largely the same. I mean, it's different if you're in cell and gene therapy, there are different processes — but it's the same for medicines broadly — everyone needs to get faster, everyone needs to get cheaper. It's not that one company is going to do it and no one else will follow — they'll eventually follow the same approach.

Sabine Hutchison
Sure. That's true. Well — some more food for thought for me, and some things to ponder. As always, I think every conversation I have with you, I leave with my wheels spinning, thinking, Christina, what can I do now? So thank you so much — it's been a treat to have this conversation. I'm sure we could go off in lots of different directions, so we'll see if we have you back again, and maybe focus on a bit of a different area. And to everyone listening — Christina, if you have any suggestions on books, articles, or things you find helpful, share those with us, and we'll share them in the notes of the podcast. That'd be great — it's always good to learn from others about what's a valuable read.

Christina Busmalis
We'll do that — great. No, thank you very much, I really enjoyed the conversation, happy to talk again. And wish you a good Monday.

Sabine Hutchison
Thank you, you too — wishing you a good week ahead.

Christina Busmalis
Thanks.