[00:00:00] Brandy: Hi, and welcome to Katalyze AI, the series where we unravel the intricacies of biomanufacturing. Join us as we explore the hurdles, innovations, and visionary ideas shaping the future of this dynamic field with some of the brightest minds in the industry. Thank you for joining us on this journey. We're glad you're here.

[00:00:22] Brandy: Hey everyone, welcome to another episode of Katalyze. Today, I have Mike Burton with me. He has had a super distinguished career in biopharma Especially in drug development. Mike, welcome to the show. Thanks for joining us.

[00:00:38] Michael Burton: Thanks for having me. Thanks for having me.

[00:00:41] Brandy: Could you start by sharing a bit of your background and how it has shaped your macro perspective on the industry's evolution?

[00:00:48] Michael Burton: Yeah, sure. So, uh, though, I know I look like I'm, uh, I'm 16. I've, I've been doing this for a long time. Um, I spent about 15 years in operational roles for life sciences, pharma, CRO space, um, all the way at the sort of lowest level, working myself all the way up into managerial and director type roles. Um, and so, you know, really, really appreciate the world and sort of what's really going on there.

[00:01:17] Michael Burton: But then for the last, let's say, a dozen years or so, I've worked for software companies, 9 of which was working for Viva and I'm sure is pretty well known to most folks at this point around clinical quality and regulatory. Industry focused software solutions. Um, and then for the last three years, I've worked for a health tech startup.

[00:01:37] Michael Burton: So I think that I have a really good perspective on sort of where we were back in the nineties, transitioning all the way through the Internet age. Let's say the initial Internet based technologies, seeing the sort of real big push to the cloud in the early 2010s. Um, and then now sort of, you know, getting into even cooler and crazier stuff that I'm sure we're going to touch on here a little bit.

[00:02:00] Michael Burton: Um, and so, you know, I'm not going to classify myself as sort of a manufacturing guy, but I would classify myself as an industry guy that has seen a lot, paid attention to a lot, um, and maybe more importantly in the last decade or so is, is just talking to tons and tons and tons of customers, uh, big pharma, little pharma, little biotechs, et cetera.

[00:02:20] Michael Burton: So, uh, I think I bring that, uh, perspective to the table here.

[00:02:24] Brandy: Yeah. And, and a lot has changed over the years since you've been in the industry, right? And even most recently, yeah, I can't,

[00:02:33] Michael Burton: I

[00:02:34] Brandy: can't even imagine. I mean, can you just talk through just the increasing, increasingly complex and expense? That's come with the biopharma industry and how companies are just struggling to keep up with everything.

[00:02:50] Michael Burton: Yeah. I mean, I can remember, you know, back in the nineties, it was literally just stamping out pills. And so you'd come up with a pretty, you know, interesting compound and from a manufacturing perspective, it was pretty straightforward. I'm not diminishing the work that was done. It was, you know, wasn't difficult or complicated at the time, but it's pretty straightforward.

[00:03:10] Michael Burton: So you were stamping out pills and. put them in a bottle and ship it about. Um, and then, you know, somewhere in the last, you know, 20 years or so, we started getting the biologics and more complicated vaccines. But to be quite frank, we really got into much more complicated disease states that required much more complicated compounds to, to manufacture.

[00:03:31] Michael Burton: And so now you've got to create manufacturing plants that cost billions, um, and, and can run all day long and maybe even still only produce a liter at the end of the day. And so just that, that inflection somewhere in the middle there, I can't pinpoint exactly, but Um, just made things just so darn complicated.

[00:03:52] Michael Burton: And on top of that, you know, the pressures of trying to do things globally and, and trying to do things at a profit, et cetera, et cetera, et cetera. And so when, when manufacturing complexity started getting crazy for a million different reasons, the tech didn't sort of keep up with it. And so, and, or it just was never a good time, I guess.

[00:04:13] Michael Burton: You know, if you think about, say, clinical trials or or regulatory that have probably enjoyed a lot more technology innovation, um, you know, there were opportunities to inject new technologies, whereas with manufacturing, it's just a wheel. It's just got to keep turning. It could never sort of turn it off.

[00:04:31] Michael Burton: And so, as a result, then, um, you know, you, you only have a couple of very focused, um, software applications that get developed for that space. And as things get more complex, those older systems, as we know, if we think about, you know, systems we used or phones that we use 10 or 15 or 20 years ago. That's just not going to keep up with the world that the way it stands now today.

[00:04:55] Michael Burton: Um, and so now you've got sort of legacy applications that you're being forced to use. And, um, the old joke used to be that, like at home, we had the best stuff and then you go to work and you'd be tortured with all these bad, ugly technologies. Why can't we have all the best stuff at work too? Um, that definitely has changed.

[00:05:13] Michael Burton: There is definitely a lot more, but, um, with, with manufacturing specifically. You've seen a lot of like mergers and acquisitions, et cetera, because to create a net new plan is just so darn expensive and so darn hard. And so now what you're left with is just disparate data. You've got many different sideload systems from many different vendors.

[00:05:34] Michael Burton: Um, and so what then many, um, sort of, you know, manufacturers would have to do then is do integrations. Um, and so as soon as you even say that word, in many cases, people are like, cringe, like, Oh God, integration that, you know, it's going to take a year and cost me a million dollars and whatnot. And so a lot of custom things, a lot of data hubs, a lot of data warehouses, et cetera.

[00:05:56] Michael Burton: And, but unfortunately in many cases, the data is stale. It's old. It's not even the latest, greatest, and or. The applications still keep getting upgraded and then the integration falls apart. And so again, you know, you think about these poor folks out there in the world trying to do this stuff with just, you know, old tools, old software, old data, and trying to just do the best that they can.

[00:06:19] Michael Burton: In many cases, just gut. You know, like, Hey, we think that this is going to be something we have to do or something we have to do. And so, you know, that's a, that's a shame. And so it's, it's not a real pretty picture out there in many cases. I hear, I've heard many nightmare stories about manufacturing and you just sort of figure it out.

[00:06:37] Brandy: Can you share some of those stories or one of those stories? Yeah,

[00:06:39] Michael Burton: I heard a story recently that was, I was talking to someone about their manufacturing capabilities, um, for a very sort of, I won't go into great detail, but you know, pretty complicated product. And they were telling me that. They have to produce a couple pieces of the product in a very specialized plant in Europe.

[00:07:01] Michael Burton: And this particular product has to be kept refrigerated. So they'll, they'll build up a couple of the pieces, if you will. I'm not trying to give too many details here. And then ship it cold back to the United States to a manufacturing plant that then completes. some additional manufacturing steps and fill and finish there, and then ships it cold back to Europe so that it can then be done some additional pieces to it and then ship it out globally from there.

[00:07:27] Michael Burton: That's today. That's today. That's not like, you know, you know, that wasn't 20 years ago.

[00:07:32] Brandy: Yeah, this

[00:07:33] Michael Burton: was, this was, this was recent. And so That just goes to show that they don't have the capabilities and, and they'll look, you know, full disclosure, that's not a big giant company, but just goes to show that even in today's, you know, pretty, pretty cutting edge world that we still got lots of legacy stuff.

[00:07:52] Brandy: Yeah. And I would imagine that that's a common scenario for a lot of companies. Just given what, what has been in the past and what we're working with today, And the expense that's involved, like you said, I mean, billions of dollars to create a new manufacturing facility. Well, who can do that and yeah,

[00:08:12] Michael Burton: well, that's interesting.

[00:08:14] Michael Burton: That's given rise to a lot of contract manufacturing, um, which is a sort of bustling, you know, nice big industry. Um, but then you lose some control over certain things. I'm not saying that contract manufacturing is bad. It definitely serves a need, um, for certain things, but, um, in many cases, um, manufacturers want to be in control more and, or distribute faster.

[00:08:37] Michael Burton: Or do some interesting things, et cetera. Um, so look again, this is all, it's all opportunity. Um, but, um, yeah, it's, it's, Just for a million different reasons, manufacturing has sort of been the one left out, if you will, clinical, clinical and regulatory have sort of enjoyed a lot more innovation and manufacturing has sort of been left behind a little bit.

[00:08:59] Michael Burton: So it's an interesting area for sure.

[00:09:02] Brandy: Yeah. And something you mentioned too, is, is as this data is getting integrated and it's a, it's a timestamp, right? It's a, it's not necessarily real time data. It's at the moment, but as we're getting more. AI technologies that are being incorporated into the space.

[00:09:22] Brandy: What, what does this look like? Is, is there going to be that ability for that real time data to make real time decisions? How can, how can this be realized in the manufacturing setting?

[00:09:33] Michael Burton: Yeah, no. So, you know, AI is obviously the, all the buzz and all the rage. It probably can't go even a day without seeing or hearing about it.

[00:09:41] Michael Burton: But, um, First off, quick plug. If you're not playing with AI every single day on your own, um, you're missing out. These are new tools that we should all be leveraging. This is just like the internet. This is not something to be, um, scared of or worried about, et cetera. Um, and, and contrary to probably most, uh, headlines today, it's not taking everybody's job and, and, and sort of, uh, taking over the world just yet.

[00:10:05] Michael Burton: This is an inflection. This is an inflection point. And, and those are, those are cool. You know, in this case, we've probably now. I've sort of lived through, let's say, maybe Two or three of these now, and you know, it's going to fundamentally change everything. Um, unfortunately this is being painted in some cases is maybe a little scary, et cetera.

[00:10:25] Michael Burton: And so some people may be shying away from it or, you know, in a highly regulated space, like pharma, like sciences. People are just very cautious about this, which is understandable. Again, I, I don't want to take, uh, uh, a medication that had an algorithm make a decision about something that was wrong, perhaps for whatever reason.

[00:10:44] Michael Burton: Um, but, but this represents a big opportunity and, and, and from my perspective, why it represents such a big opportunity is because There are certain things today that AI is very, very good at, I completely understand and appreciate certain of the hallucinations and other things that people have probably seen or it can't do simple math just yet, etc.

[00:11:04] Michael Burton: But in this case, what we're just talking about is gigantic data sets, huge, huge, huge amounts of data. Being produced from many, many, many different systems, uh, globally at scale. And so when you get that much data and you don't have a really good place to look at it all or analyze it all, that creates lots of challenges.

[00:11:28] Michael Burton: So humans could go and play and sort and filter and run queries on the data and look for certain things. And I'm sure that that as has served well. But we're at a place now where, um, these tools are available. And so you'd be sort of silly not to use them. And so as long as it's constrained to something very specific, and in this case, maybe just saying, Hey, look at this big giant set of data across 10 different systems and look for a pattern, look for a red flag, look for something like that, to me, that is just, you know, just makes sense.

[00:12:01] Michael Burton: And unfortunately, good, bad or otherwise, we, we live in a very competitive, complex world. And so, um, I can guarantee you that most are looking at this now and playing around with it, but they're trying to figure out, you know, what's the right or wrong way to do it. Um, and or, um, making sure it's not making decisions that a human is not going to, you know, continue to reinforce, et cetera.

[00:12:24] Michael Burton: But, um, these tools are very, very, very powerful. Again, I'm trying to use it literally every single day for something. Just test it, play with it. Um, don't think of this as someone like taking over the task perhaps, but as an assistant to you, sort of like, uh, uh, a friend or a colleague or someone that's there to sort of help mentor you or guide you or say, Hey, look here, et cetera.

[00:12:49] Michael Burton: Um, so again, Well, you know, change is hard and I appreciate if you've been, you know, in some of these spaces for a long, long time, you sort of want to not necessarily move forward with these things, but honestly, I think you're silly if you're not at least exploring the use of some of this stuff today.

[00:13:06] Brandy: Well, and you can see the practical application of this, right? You just talked about how people, human beings are looking at this data, trying to find correlations, which takes a lot of time. And let's say you have an AI co pilot that can just help. And assist, right? Like, you know what you're looking for.

[00:13:26] Brandy: So you're feeding, you're feeding it that information. So again, you have to, you have to be knowledgeable enough to be able to do that, but then it's assisting you in finding those correlations so that you're not spending days, weeks, months. Connecting all of that data,

[00:13:44] Michael Burton: uh, well, may maybe even or, or even missing something, to be quite honest with you.

[00:13:48] Michael Burton: Yes. Again, you know, we, we, we, I think a lot about pattern recognition. I think a lot about trying to identify red flags, but, and again, you, you know, human is probably smart enough to know how to sort and filter certain things. I know I, I love playing with data myself and looking for things, but I can't always find everything.

[00:14:06] Michael Burton: Mm-Hmm. . Um, and so what you could see though is that. In these very, very complicated products that we're looking to get the yield that we're, we're trying to achieve is still a challenge. We're still not seeing 100 percent yields across a lot of things. And going back to that very initial example that I was giving, where a whole manufacturing plant might work all day long to get a leader, even just getting 5 percent or 10 percent better at the end of the day makes a huge, huge difference.

[00:14:38] Michael Burton: And so, in this case. I start thinking about, you know, uh, you know, baking it's sort of funny sort of analogy to use here. I don't mean to minimize anything around manufacturing, but, um, again, I'm not a baker, but I'm told that if you measure the ingredients by volume versus weight, that you get a pretty different, you know, uh, outcome.

[00:15:01] Michael Burton: And so in this case, if we're able to Um, you know, serve up the recipe and doing it in a way that we're taking all the input and all the issues and challenges that were identified by AI, um, or algorithms, et cetera, um, to identify where challenges have been in the process steps, um, and can be very, very conscious of that.

[00:15:26] Michael Burton: Uh, then I think that that can have a huge, huge difference and the unlock. Is going to be a huge amount of value. You're able to produce more, um, maybe even faster. Perhaps, um, look to streamline certain steps. or make better decisions. To be quite honest, I think that's what we really want. Um, we don't want necessarily more choice.

[00:15:49] Michael Burton: We want less choice. We want to be able to say, Hey, what's the right thing to do here? I imagine in some of these cases, there's probably many different decisions that you could make. And so again, in this very, very complex, competitive world, being able to produce even a little bit more from your, you know, your manufacturing is going to unlock a tremendous amount of value.

[00:16:11] Michael Burton: And so again, it should be something that, that, that all folks in manufacturing are really thinking about today.

[00:16:17] Brandy: And this is, this is where Katalyze comes in to play, correct? Yeah, exactly. This is what they're doing.

[00:16:22] Michael Burton: Yeah, pretty wild. And, and, you know, again, we're not, uh, you know, uh, I'm not here to sort of say, like, you know, I've used this myself, et cetera, but I've seen a lot of what it can do, bring all the data together in one place, um, and really help to identify and be that copilot to you to say, hey, If you want to look to maximize your yield, these are the things that you need to do.

[00:16:47] Michael Burton: Now, each of these individual things might even be very complicated on their own. And what you'd be able to do then is prioritize them. Um, you know, they sort of say, maybe you could do one or two or three big things in a year. Um, but if you don't know, if you're not focused on the right things and or in the most value, um, this will give you the data to be able to say, Hey, if you were going to do one thing.

[00:17:12] Michael Burton: This would be the area in the manufacturing process where you should focus, where you should really look to maximize. And so that's that's powerful. And I think in many cases, people are sort of begging for these kind of tools, um, and for a million different reasons, just maybe not able to sort of pivot quickly enough or explore these things.

[00:17:36] Michael Burton: But again, I would, I would implore you to, um, you know, be playing with AI all the time. And we're looking for solutions like this to help you out because this is, this is clearly the next wave.

[00:17:47] Brandy: Yeah. And, and of all things, Mike, you talked about how there's some fear and obvious, obviously there's regulation in this space, but yield optimization seems like a good place to begin.

[00:18:00] Michael Burton: Oh, that's actually a great point. You would sort of say, look, there's probably millions of different use cases here. I'm sure that all kinds of people are thinking about all these, um, things that can be, um, sort of done with these kinds of technologies automation wise. Yeah. You're absolutely right. I think I'm not going to say that's the exact, always first place to start, but it seems to me when it's a straight up data play like this is and identifying and starting out, it would clearly be somewhere that would be straightforward, um, and could Produce a pretty dramatic effect.

[00:18:38] Michael Burton: It would seem like certainly one of the early use cases to be exploring, especially as we continue to learn more and more about this stuff. The other crazy thing that we haven't really touched on is that the rate of innovation here now, too, has been just a Astounding. Um, you know, again, when these early models came out, you know, uh, maybe two years ago or so, there was clearly gaps.

[00:19:01] Michael Burton: But now there's been new models introduced even in the last month or so that are now able to do all kinds of things where, um, they're teaching it, making it better and better and better. Um, the amount of resources and people that are focused on making these AI products better and better is just amazing.

[00:19:19] Michael Burton: And so, yeah. I even sometimes struggle to keep up with it. I feel like I read the news every single day about this stuff. I love it. Cause I think it's really cool. Um, and every day I'm learning about some other new piece of it. Um, so yeah, look, this, this stuff's really cool and, uh, it's definitely something that people should be looking at.

[00:19:37] Brandy: Yeah. And you're in a very unique position too, because you're advising a lot of different organizations. So you get to see the development. Faster than anyone else in the in the industry.

[00:19:49] Michael Burton: Yeah, I am. I've sort of been advising a bunch of sort of AI startups.

[00:19:55] Brandy: Yeah,

[00:19:55] Michael Burton: and well, a amazing just to see all the new companies that are being created as a result of this.

[00:20:01] Michael Burton: And that's how you could also sort of feel like this is a major transition when. When folks are now putting money and time behind these things, um, yeah, it's wild things that weren't even like conceivable, even a few years ago. And so, yeah, I am seeing a lot of different cool AI startups for sure. Um, and yeah, I could definitely tell you that this is.

[00:20:24] Michael Burton: Of the ones that I'm sort of been aware of, et cetera. This is one that is certainly one of the most exciting that I've seen for sure, because again, it's so clear and straightforward. And, and, and again, from a value perspective, tremendous.

[00:20:39] Brandy: Yeah. Talking about Katalyze.

[00:20:41] Michael Burton: Yeah, exactly. Yeah, exactly. Yeah.

[00:20:43] Brandy: Yeah.

[00:20:43] Brandy: Well, this has been great, Mike. Thank you so much for sharing your, your knowledge, your expertise. I guess if you were to leave this conversation, leave the listeners with a piece of advice, I mean, you've given this already, but everyone should be playing around with AI, but in the industry, what, what are you, what do you recommend that people do pay attention to any piece of knowledge that you'd like to, to share?

[00:21:07] Brandy: To drop or advice you'd like to share.

[00:21:09] Michael Burton: Yeah. That the first one would have been just play, right? You can't, you can't mess anything up here. I'm and don't think about asking big giant questions of it. When you use these sort of, you know, chat to your other models, uh, dig deeper, have a conversation with it.

[00:21:25] Michael Burton: So you can certainly say, Hey, you know, write me an essay, whatever. But then when it comes back, it's like, why did you do it this way? Or. Tell me one thing that I missed that we should have been talking about, you know, dig deeper into it, play with it, have a conversation with it. Um, and I think you'll be astounded at the level that it's at today.

[00:21:44] Michael Burton: Um, and then the second thing is, if you're an organization out there sort of thinking about this, I'm sure I'm hearing about many that are having like a AI committee or some kind of a, a board that's looking at a lot of these things. You got to start with something. Just start with some simple use case.

[00:22:01] Michael Burton: Um, to start getting familiar with it, seeing how it's looking, et cetera. Certainly pay attention to what the regulators are saying. So look for the lowest risk type of use case. But I think you got to just start with something. Um, and then just take one step, one step, one step. And I think you'll be amazed.

[00:22:17] Michael Burton: At the amount of value that these new tools can bring.

[00:22:21] Brandy: Incredible. Thank you so much, Mike. Really appreciate your time.

[00:22:24] Michael Burton: Thank you.

[00:22:28] Brandy: This series is brought to you by Katalyze AI. The leader in equipping manufacturers with cutting edge AI tools. Katalyze AI redefines manufacturing by digitizing, interlinking, and enriching vital data. Implementing plug and play AI modules for SWIFT results and developing personalized AI solutions. If you are interested or want to know more head on over to Katalyzeai.

[00:22:54] Brandy: com to find out more information.