How AI is Shaping the Future: The Rise of The Bionic Professional with Bill Wallace

How AI is Shaping the Future: The Rise of The Bionic Professional with Bill Wallace #aitechnology #futureofwork #bionicprofessional 

Step into the exciting world where AI is transforming the way we work and the very nature of our professions. Please visit our website to get more information: https://swangroup.net/ 

In this episode, I’m joined by Bill Wallace, Chief Information Officer at Intercept Pharmaceuticals, as we explore how AI is reshaping industries, and the rise of the bionic professional. Together, we dive deep into how biotechnology, AI, and professionals can work hand-in-hand to create smarter, more efficient workplaces.

From the advancements in biotechnology to AI’s integration into daily business operations, discover the potential for professionals to become bionic, enhanced by technology for maximum productivity and innovation. Bill shares his insights on AI’s current role in business, and what the future holds for the professionals who embrace these changes.

Unlock valuable perspectives and stay ahead of the curve with this in-depth conversation about the future of work and AI’s growing impact on our professional lives.

Links from this episode:

• Get to know more about Bill Wallace: https://www.linkedin.com/in/williamwallace2 
• Learn more about Intercept Pharmaceuticals:https://www.interceptpharma.com

🔔𝐃𝐨𝐧'𝐭 𝐟𝐨𝐫𝐠𝐞𝐭 𝐭𝐨 𝐬𝐮𝐛𝐬𝐜𝐫𝐢𝐛𝐞 𝐭𝐨 𝐨𝐮𝐫 𝐜𝐡𝐚𝐧𝐧𝐞𝐥 𝐟𝐨𝐫 𝐦𝐨𝐫𝐞 𝐮𝐩𝐝𝐚𝐭𝐞𝐬.
https://www.youtube.com/@Biotech_Bytes/?sub_confirmation=1 

🔗 Stay Connected With Us.

LinkedIn: https://www.linkedin.com/company/the-swan-group/ 
Website:  https://swangroup.net/   

=============================

🎬Suggested videos for you:

▶️ https://www.youtube.com/watch?v=ZWXIIe66-kI 
▶️ https://www.youtube.com/watch?v=T6EzJ1F_6pg 
▶️ https://www.youtube.com/watch?v=be8szNVFrNk 
▶️ https://www.youtube.com/watch?v=_Be6WEEy2JM 
▶️ https://www.youtube.com/watch?v=mqpB3pGywkU 

=============================

#aitechnology #futureofwork #bionicprofessional #aiworkforce #aiimpact #techinterviews

How AI is Shaping the Future: The Rise of The Bionic Professional with Bill Wallace

Chapters

• 1:45 About Bill Wallace
• 6:23 AI use cases: marketing, service, development
• 11:13 AI’s role in business productivity vs. individual productivity
• 17:34 The real ROI of shortening clinical trial timelines
• 27:52 How data structure impacts AI effectiveness
• 32:53 Guardrails, governance, and security with AI
• 38:00 Why collaboration still matters in an AI-first world

Transcript

Steve Swan [00:00:00]:
Next on Biotech Bytes, join me, your host, Steve Swan for an awesome conversation with Bill Wallace, the CIO of Intercept Pharmaceuticals, on the evolving role of AI within biotechnology companies and the rise of the bionic professional. Yes, bionic professional. Tune in. Welcome to Biotech Bytes. I'm your host, Steve Swan, where we chat with IT leaders within biotech. Today we have a treat. We have our first repeat customer, Bill Wallace, head of IT at Intercept. We last spoke with Bill, I think what, about a year and a half ago or something?

Bill Wallace [00:00:39]:
Bill? Yeah, January of 24, the first biotech Bites, if anybody wants to go back and take a look.

Steve Swan [00:00:48]:
Yeah, exactly. And so this is great. Want to to, you know, sync back up with Bill, do a little update, hear what's going on, hear what he thinks has changed in the last year and a half. And as always for our listeners, if you like what you see and or hear, don't forget to follow us on Google, Spotify, Apple and like and like us. So Bill, welcome. Thanks for joining us again.

Bill Wallace [00:01:13]:
Thank you. Great to be here, Steve.

Steve Swan [00:01:16]:
So, so as you know, as we've done in the past, you know, in case somebody's watching you here for the first time, they may go back and watch the other one. But just to give folks a brief introduction on you that may not have seen the last one, can you give me a quick, you know, I don't know, minute or so on on your background and what's led you to be the head of IT here at Intercept?

Bill Wallace [00:01:37]:
Sure. So I have a varied background. My first 10 years, I actually was not in technology at all. I was on the business side, first in finance for five years and then in marketing for five years. And then something really wild and crazy called the World Wide Web came along and that started my transition to technology, which is where I've been ever since. And although I haven't worked in several industries, most recently I spent the last several decades in pharmaceutical and biotech. And along that time I've worked for some of the largest companies, Pfizer, Midsize, and more recently have been CIO at several organizations, some of which were building their teams. I had to build the teams from scratch because they were pre commercial looking to launch their first product.

Bill Wallace [00:02:27]:
And so I was their first hire in it in several cases when they were under 100 employees and about to grow very rapidly. So now I've been with Intercept, in December, it'll be four years. And I'm very eager to talk about how AI has changed even in the year and a half. Plus that We've talked, let's get into it, right.

Steve Swan [00:02:51]:
So last we spoke, we talked about AI, right? And like you said, it's, it's changed a lot. I mean, I think probably every couple weeks you can say that things have completely morphed. So since the last year and a half, tell us, give, give us an update on how you see it, what's changed, and maybe where you think we're heading. I know that's going to be a mouthful, but, you know, let's, let's, let's dive right into it.

Bill Wallace [00:03:13]:
Well, I would say that the general capabilities have changed quite a bit, mostly for the positive. And at the same time, just about everybody is using AI in some way, shape or form, whether they're using it for personal or professional. And the, there's still a lot of learning going on and some people know it better than others and are more interested than others. But it is one of those things where you pretty much can't have any conversation about business and technology without having a conversation about AI. It is truly where the Internet was in the early 90s in terms of business and technology starting to partner and collaborate together to figure out how it can be utilized by the organization. And I wish I could take credit for this, but I have heard the term and I think it's a very good one. Bionic professional. Just like the Bionic man and the Bionic woman in the show in the early 70s, the AI.

Bill Wallace [00:04:30]:
And if someone is able to really utilize AI within their business processes, it absolutely can be of a great help to them and make people more efficient, more effective, sometimes more timely. And at the same time, it's still very important to have the knowledge that the individuals, the individual professionals have. Because if there's one thing we've all heard about the AI hallucinations, AI itself can pull a lot of data together and you teach it so that it pulls it in certain ways, in certain aspects. But ultimately, for a lot of the best business insights, you still need a professional who has the experience in their given space to really be able to look at what's coming in and figure out what is truly insightful and what might be going down a bit of a blind path. Because AI itself, you know, the key part being the artificial part of intelligence, it doesn't in all cases put things together in the exact way that, that make logical sense. And just about everybody has seen this when they've used AI for their prompts, searching on the Internet or different things. But it is something that is very important to understand and have that business background, skills and experience when you're utilizing AI from a business perspective. Which is why I think bionic professional is a great term.

Steve Swan [00:06:15]:
Yeah, the Bionic professional. I hadn't heard that one. I hadn't thought of that. You know, last week I was chatting with a gentleman who's a CIO of a, you know, probably about a billion and a half dollar company. And he really focuses and he always has focused on, you know, roi. What's going to be, what can I point at, you know, for any technology that I'm utilizing or putting in place or whatever, you know, to the bottom line. And as he was looking at it, I asked him, I said, so tell me, what do you, you know, when you're looking at AI, where do you think your biggest bang for the buck comes with AI? And he, he, without even knowing he was saying what you just said, Bionic Professional. If I list all three of his, his areas, it's the Bionic professional.

Steve Swan [00:06:56]:
Number one was content marketing. Marcom, you know, marketing content. You know, he's like, there's great return on that. Number two, customer service, you know, helping those customer service reps to become just again bionic, you know, being able to root calls and being able to do things that they couldn't otherwise do. And then finally the third one was truly the Bionic Professional where it's for developers, it's going to develop all their code, it's going to write their code. Now the developers have to know the code in order to allow the machine or AI to write it because they do have to still do code review but they become, they, they climb up the ladder and they become more architects. Right. So they're utilizing AI to again do the mundane piece of, of the code but now they can focus on again business architecture and code review.

Steve Swan [00:07:43]:
So truly like you just said, I like that Bionic Professional, you know.

Bill Wallace [00:07:48]:
Absolutely. And when you, when you think about something, you mentioned the code and the developers. And so we also have some coders who have been leveraging some AI and overall it has made them more efficient. But getting back to the Bionic Professional piece is you can't just put a couple of prompts in and then have the AI create your code and you're done. They find that AI sometimes creates inefficient code. It, overall it makes them more efficient because a lot of the easy stuff the AI can do very quickly as opposed to someone having to use ten finger magic and type. But by the same token you still need that developer coding professional who really understands all the logic streams, because AI doesn't as, as much as there's been a lot of training, it doesn't always fully understand that and doesn't always fully understand how the code needs to work with those business processes. And.

Bill Wallace [00:08:48]:
But putting the two together gets you that, that Bionic Professional and the customer service and marketing content. I mean, you've probably seen where people are now using AI to create animation, videos and all kinds of different things. Things that used to take a very long time and a lot of computing horsepower and deep skills and there's still skills that are needed, but a lot of the computing horsepower is now being done by AI. And this really gets to. I think of AI as two types of efficiency. One is individual productivity and the other is truly that business productivity. And that's from my perspective, where most of the business, the Bionic Professional comes in. So if you think about personal productivity, probably at this point, just about everybody watching this has used AI.

Bill Wallace [00:09:49]:
Okay, I wrote an email or I wrote a letter, let's see if you can make it sound better. And I call that individual productivity or hey, look across my calendar, see when somebody can. A group of people can get together for a meeting and again, depending on what kind of AI you have going on. But that kind of individual productivity helps at the individual level, but it doesn't really dramatically impact overall business processes. Nothing wrong with developers being more efficient with their coding. Nothing wrong with people who have to send emails to a bunch of people using AI to be able to craft those emails quicker and faster and get that done more quickly. Individual is great. But really where that bionic professional, I think comes into play and where the collaboration between business and it is so important is when you look at the broader business processes and to your point about focusing on roi, that's where I think that you can see that the focusing on those business processes can potentially get you the ROI.

Bill Wallace [00:11:05]:
I'm not saying that if you have 20 developers and all of them start utilizing AI at some point in the future, you can get the same amount of work done with 19. But by the same token, usually then once they're more productive, people think about all that much more that they want to get done. But from a business process perspective, if you put AI in the right places, it truly can change those business processes. And that is where you can see some true roi. I liken it back to automobile factories, right? They have workers. Well, in the 80s for the first time, they had these wild and crazy things coming in called industrial robots. And they came in and it took about 20 years. But if you think about where it was in the beginning of the 80s when they first brought in industrial robots to where it is today, it has meant a wholesale change on the assembly line level.

Bill Wallace [00:12:10]:
We still have thousands of people in automobile plants, but there's less thousands than there used to be. At the same time, those plants are more efficient because the industrial robotic portions have been placed in those areas where they're most efficient. And you have the human workers working where they are most efficient. And I see a lot of that in conversations with business use cases for AI, with our, with our business stakeholders and colleagues.

Steve Swan [00:12:43]:
That's actually what this, this individual mentioned to me. You know, almost different terms, right? But you know, when he talked about, you know, I look at the ROI and I look at where it's going to flow, you know, he said, listen, if accounting comes to me and says I'm going to use him and I'm going to reduce our headcount by one, that doesn't get me excited. But like you just talked about, if we put it in for, for business processes and, and it helps ROI and a big, bigger scale, that gets us excited again. The marcom, the, the, the customer service, you know, that kind of thing, maybe the code development. But he said, one head count, that, that's great. You can have him reviewing contracts and that's one person, but that's not going to, that's not moving the needle in his terms, you know. So I had another gentleman that I was chatting with and he calls me, he says, you know, Steve, everybody tells me this, this guy does a lot of work in manufacturing, mes kind of stuff, right? And he said, you know, a lot of people are talking to me about AI and AI is coming from my job and all this stuff. Maybe it is, maybe it isn't.

Steve Swan [00:13:44]:
But I gotta tell you, he said, I've been doing this a long time and what I do know about AI is how specific we have to be when we're dialing IT in and training those models so we know exactly what we're solving for. He said, I've been an IT professional for a long time. Every time I sit down with an IT with a, with a business group, they don't know what they want. So how are they going to plug into AI exactly what they want and have AI solve it? Because AI only can solve what you ask IT to solve. So most of these business folks, when we work with them, we've got to tweak and work with them to tweak what they are really looking for, sometimes they don't know if A, what's possible and B, what they really want. We got to tease it out. AI is not doing any of that. So we said, I'm really not too worried about it, at least nowhere in the near future because that means that business suddenly will change in the respect that it's going to know exactly what it wants from minute one and AI is going to go solve that.

Steve Swan [00:14:38]:
So I, I, you know, I heard that and I thought about it for a second. I said, you know, that's kind of age old. I've been hearing that same sort of conversation or rhetoric. When we, when we got rid of the business analysts 20 years ago, when we, you know, again, different iterations of business partners, right, we, we kind of talked about that. You know, business partners has become such a huge thing nowadays. It, it's getting bigger and bigger. IT business partners. And, and that's the reason why, because you know, it's, a lot of the businesses has always just needed somebody to help them navigate through what's possible with IT and what do they really, really, really need? You know, Absolutely.

Steve Swan [00:15:19]:
I, I, I've seen that a lot, you know, and I think we're going to continue to see that. And that's why this guy said to me, yeah, I'm not too worried about my job, you know. So what are your thoughts on, on that? Do you agree, disagree to where, where would you fall on that?

Bill Wallace [00:15:30]:
I would absolutely agree. When you, I mean if you think about what everyone talks about with AI oh, we need the data, we need the data. And I, I'll talk a little bit about data and data architecture in a moment. But if you think back to okay, we need reports and we need reporting. Technology has worked with the business for many years to build reporting systems to get the KPIs and the metrics they need. And one of the key things you want to do from a technology perspective is when you're getting together with the business and trying to define those reports and those key performance indicators, you really need to help often do the translation between the baseline data and what it is that the business is looking for. Because very often, oh well, I want to know how to help the sales team get to that next best action. Okay, great.

Bill Wallace [00:16:21]:
Well, what data and information around the market is going to help provide that list? Oh yeah, we got to talk about that. And you know, what, what data can be done and how should it be put together. And I think a lot of the business use case for AI to get to those bionic professionals absolutely requires the same thing. Because people talk about the hallucinations of AI. Well, that's because the AI is pulling whatever data it has access to. And again, it is artificial part of intelligence. And so it sometimes pull it together in ways that don't make sense for the specifics of what someone's doing. That's where a lot of very good coders are working to train AI models because you need to put those constraints about how the data relates to it together to each other and how the business user is going to go in and try and access that data through an AI engine.

Bill Wallace [00:17:26]:
And those are key elements and aspects and that is essential. And you know, from an ROI perspective, sure, everybody's worried about jobs being lost. And I'm not saying that as things are more efficient that that won't happen. I look at it in terms of, again, if you use the example of containerized shipping, we all know what containerized shipping is. Well, it came in in the 1950s. Before that they didn't have containers. When it did come in, there were several hundred thousand dock workers across the US who no longer had a job because they didn't have to do cargo nets every time they unloaded a ship. But the containerization has made things more effective.

Bill Wallace [00:18:15]:
And the dock workers today who are up on those cranes loading and unloading those ships are in, you know, really nice six figure salaries. And so ultimately that productivity benefited the workers who were there. And I think that when we look at AI, it's going to be similar. And when people talk about ROI and they think about, well, what does that mean for headcount? I think that's, that's really focusing on the wrong thing. We're doing some very, very interesting stuff that for competitive reasons I can't get into very specifics, but we're doing some very interesting stuff on the clinical space. And if you think about it and you say, okay, we have a clinical trial and that clinical trial is several years long and then we get the data and then we have to lock the database. And of course it all has to be validated through the whole process. That's, you know, absolutely essential.

Bill Wallace [00:19:12]:
And, but as they go through it, at the end of the day, the product ideally gets submitted and assuming approval, you know, what is the market value of that product going to be, say, 24 months on, for whatever time frame it is. And I'll make my math easy, right? If you say that 24 months on this product is going to be worth $360 million a year in revenue. Then every month that you can shorten a three to four to five year cycle for getting to the point where you can get through the trials, get the data, get it locked, and then go through the analysis, which also can use AI and get it approved. Now you're talking savings of $30 million or I'm sorry, additional revenue of $30 million for every month you save. It does not take a long time. If you can put together some AI capabilities that can shorten a trial set for one compound by even a month or two. If you can do that for two or three, you have one hell of a return on investment for AOI and no ROI rather. And no one has to think about for that return on investment for AI, what that meant for a headcount.

Bill Wallace [00:20:37]:
Right.

Steve Swan [00:20:37]:
I mean, that's huge, right? Those dollars. There was somebody I was talking to the other day and they, they try and help organizations connect their, their AI agents that are in different verticals, you know, across. And, and they were trying to, they said to me something along the lines of having the time of the clinical trial. And I said, well, that sounds pretty aggressive, but I think, you know, if, if you talk to folks about, like you just said, a month or two or three, that's still huge numbers having the time. That's, that's, I mean, I think that's way out in the future, but maybe, maybe, maybe we're getting there. I don't really know. I was talking also to an alternative data company recently and they were telling me, you know, because I, I think they're ripe for, you know, AI and utilizing AI, Right. So, and the alternative data companies take a lot of the, you know, esoteric data and feed it to the hedge funds.

Steve Swan [00:21:26]:
And then the hedge funds will make calls based on different things that they see going on at fast food outlets and they'll Compare Wendy's to McDonald's to whatever, you know, and then if this is going, if XYZ is taking place in this particular outlet and a bunch of different McDonald's outlets versus in the Wendy's, Arby's and whatever that means that there's going to be more revenue coming into those outlets and then that Stock will be 2, 3% higher in six months, whatever. You know, I'm just making the scenario up and, and I was talking to this guy and you know, he said, we're busier than ever. The hedge fund spent a lot of money on us, all this other stuff. And right now, I guess I didn't even think of this. The alternative Data companies are probably very busy because of the government shutdown. BLS is probably hurting anyway. So what he said to me, though, as, as it pertains to AI, when I asked him about AI from, you know, from the alternative data perspective, he said, we can't use any of it because of what you just talked about, the artificial piece, you know, if it doesn't have enough to complete the circle, it's going to hallucinate that last little bit. And he said, we got to check everything anyway in these hedge funds, we're done.

Steve Swan [00:22:33]:
We're done. If they find an error and they're not making, you know, they're going to lose millions. So we've got it. Everything's human, every single piece of it, you know, because if we let you know, if it just completes a little bit of that circle, we're in trouble. And it's, we don't know if that's real or not. So he said, it's getting there. You know, we like it. But right now, the amount of money that's on the line for the trades that are being done and the amount of money that they spend on us, we just can't, we can't dip into it.

Steve Swan [00:23:00]:
Which I thought was interesting because I thought that something like that would be real ripe. But I get it, I understand. You know, so it's, you know, we're getting there, right?

Bill Wallace [00:23:09]:
Yeah. And, and when we're talking about AI for, for business processes, embedding AI in business processes, you hit on one of the absolute key points, which is at this point, and again, pharma and biotech, right? I mean, if you're just doing a search engine and you're Google, it's okay to say this is AI and they can make mistakes and everybody's got to look and see. But when you're in our space, it's a bit different. And so where we're finding efficiencies from a business process perspective, it still is a matter of, at the end, you still have a bionic professional who was doing the review and the final edits and making sure that there wasn't that hallucination at the end, or if there was that, that it, it can be fixed before it, it goes out. And I think that is where, again, right. There's still a lot that AI can do when it's done well and appropriately to make things go faster. Getting back to that example about coders and having IT write some code. But at the end of the day, those developers or those business stakeholders who have that professional Knowledge and understand their business processes and what's expected.

Bill Wallace [00:24:29]:
They need to be the ones to look at that output and make sure that everything is what it is so they can gain. The process, can gain a fit deficiencies. But by the same token, it doesn't work by itself. I had a friend who is an attorney and they were going to court and they were a defendant for a company and the plaintiff who was suing, they put in their brief to the judge and they did not edit and review that before they sent it to the judge. Well, the AI manufactured court cases made them up. And when the judge realized that basically he was being given a load of garbage, let's just say it did not work out very well for the plaintiff because everything they had submitted wound up getting thrown out. So similarly, in the pharma space. Right.

Bill Wallace [00:25:36]:
Again, AI is doing, we're doing some amazing things and it's only going to get better in terms of helping the bionic professionals to utilize AI. But that professional side of the bionic is absolutely essential to make sure that the, that the process works as intended.

Steve Swan [00:25:57]:
Yeah, we can't have those mistakes because, yeah, a court case gets thrown out or, yeah, you know, whatever, someone doesn't buy that cheeseburger. But we're dealing with human lives. Our mistakes are literally fatal. Right. If, if something like that happens. And in the way I understand it, you may know more about this than me. I don't know any more than I'm about to say. But I think the FDA now has a, an AI sort of ARM or an AI division.

Steve Swan [00:26:20]:
Right. So they're, they're, they're hitting that head on. They have to.

Bill Wallace [00:26:24]:
Right. Because it's here. So. Absolutely. And my understanding is they're looking at potential for using AI for some of the review of the submissions. And again, from a portion of that process, I expect that AI will be able to work very, very well for them. But it is important. You can't, again, you can't just turn it loose and then say, oh, whatever it comes up with is the answer.

Bill Wallace [00:26:48]:
You need to have those professionals, whether those, the professionals of the FDA or, or the professionals in pharma and biotech who are presenting to the fda.

Steve Swan [00:26:57]:
And I love that bionic professional. I'm stealing that. And I'll, I'll footnote you on everything when I.

Bill Wallace [00:27:03]:
Well, I love that. That's awesome. To be fair, I stole it myself because I really liked it when I heard it too. So perfect.

Steve Swan [00:27:12]:
It's so appropriate for everything we talk about. You know, I hadn't heard or read about that yet, so I like it now. So earlier, you know, and, and when we talked a year and a half ago, we talked a lot about data, right? And everybody's always talking about data because data is the gasoline that has to be pristine, that goes into these engines, right, these, these LLMs and, and, you know, so on and so forth. You, you mentioned a minute ago that you wanted to touch on some data. Tell me about that. What's what, what have you, what's new or what's different about from a year and a half ago when we spoke about, about data that goes through AI.

Bill Wallace [00:27:44]:
You know, I like to say, I've said this for a long time, data flows like water. It flows in whatever way someone wants it to flow and it can wander in various ways. And when it, early on, again, when we were first talking, I had a number of consultants who came in and oh, we're going to build this big data lake for AI. We're going to pull in the data, build it, and they will come, the business will come and figure out how they can utilize the data, data and AI and, and do things better. And don't get me wrong, I think the, the movie Field of Dreams was a great movie, but when it comes to whether it's building it for reporting structures or building it for AI, honestly at that point I didn't, I didn't return any more of their calls because I don't believe that that's going to be helpful. Just like AI, you know, when you're doing reporting, if you don't have good quality, quality data, people lose, you know, lose the ability to believe in the quality of the reports they're getting and what they're seeing and then business insight softers. And similarly with AI, I like to think of it more. If you think about, if you think about water, right, and you think about the reservoir system for a large city, you don't just have one big thing.

Bill Wallace [00:29:11]:
You have multiple, you have multiple rivers and multiple streams that feed into ultimately usually multiple reservoirs. And then there is a very structured process to take that data, make sure it's pure, make sure it's got what it needs, and then to flow it through to the endpoints so that all the people can use it and the industries can use it. And I like to think of the evolving AI architecture and a lot of it is data architecture as a very similar piece. I mean, today when you think of, oh, we need to do, we need to do business reporting, you don't have a clinical data that is in the same place as commercial data that is in the same place as your financial data. You have individual data environments with specific rules, specific rules in terms of how the data is processed as well as who can see it. And I see a lot of that as kind of the future for AI. And honestly some of what we've been putting together in our, in our AI that we're working with our business stakeholders on, because if you've got commercial data, you certainly don't want commercial, you don't, you certainly don't need or want clinical data coming into there because then something that should not be released may get out to somebody or be used in an inappropriate way. And similarly, if it's clinical data, obviously validated only, you're certainly not going to bring in commercial finance.

Bill Wallace [00:30:43]:
Why would you? Because that's not what you need to do with that. And so I like to call it, you know, I used to call it data guardrails, now it's data and AI guardrails. And you have those data spaces and understanding the specific business use cases and how that data is going to be utilized. Because even within the clinical and the R and D space and we again going back to clinical trials, even within that, there's the clinical trial information, there's a variety of things that you need and can do with that base clinical data, all the sites reporting in, but then there's also the statistical analysis. And even though ultimately it's all data from the clinical side, your statistical analysis data sets and the statistical tools you're using is in a different area than your baseline clinical data environment. And so really from an AI perspective, where we've seen it be most successful for the, again, the business processes is when you fundamentally understand what kind of data each of those business processes needs and uses. And then look at how you can teach an AI model while at the same time restricting what it can access to what the business users use today. Because that dramatically minimizes the amount of hallucination.

Bill Wallace [00:32:16]:
And even where there is hallucination, it just may be a little bit different ultimate result set without going completely crazy and being something where someone says, oh, this makes no sense whatsoever.

Steve Swan [00:32:29]:
Right, right, yeah. So where does all of this fall into? Security slash governance? I mean, you can't, like you just said, guardrails. You got to have rules in place. You gotta be thinking about all that right there with what you were just talking about, correct?

Bill Wallace [00:32:45]:
Yes, absolutely. And it's, it's one of those things where it's great to have the ChatGPT that you can go on for free and utilize on the Internet, but that's not what you. If you're going to use ChatGPT for the business, for us, that's not where you're going to use it. Because if you go in there, anyone you know, other people can look at, at what data you're putting in, what the results are coming out. Needless to say, you don't want your proprietary information in the public sector. So from a compliance perspective, I would say the easiest thing is, and we've done this for a number of our business solutions, is you bring an engine into our, in our case, our cloud environment, some people have on prem, but our cloud environment and then what is used to train that model, the configuration of that model, all of that is resides within our space. And people, there are people who are creating their own models from scratch and there's nothing wrong with that. I will tell you that we're looking to do things in a slightly different way because we're of a certain size.

Bill Wallace [00:33:56]:
And so we're bringing in known LLMs that are very good at, we've seen that they're very good at what they do. The consultants we're working with have seen they're very good at what they do. And then we bring them in and put, and we put them into our environment. For example, we have one LOM that we're using and again, I don't want to start throwing names out, but we brought it in because we talked with a consulting firm that had used it successfully for several other things, pharma and biotech clients, and they said, hey, this, this LLM is actually exceptional. The way that they baseline trained the model, it knows it really does really well with medical terms. So for us, for some of our, for some of our R and D clinical area, having that LLM made it, we actually did a test with that and two other LLMs and it blew the other two out of the park. And you know, there is really a difference. You can't just say, oh, use, you know, XLLM for everything that we do.

Bill Wallace [00:35:00]:
It just doesn't work that way. At the same time, from an architecture perspective, you can't have 20 different LLMs. So there is a bit of an artistry going on about sometimes you need different LLMs. But by the same token, we're trying to build a consistent platform even if the data is different from, for the different business use cases, we're trying to have a consistent platform so that you don't have everything be different.

Steve Swan [00:35:28]:
So these are pre canned LLMs that a consulting firm's developed, they lease them, sell them whatever you want to you guys. And you run your own internal data through that LLM.

Bill Wallace [00:35:40]:
Yes, plus, so it is an LLM. It is a base LLM that anybody can license. And then they take our data in our specific disease states, because that's a key piece if you're on the clinical side and different things, the R and D side. And then with our business, you know, giving them the background for our disease state, what are key elements that are important, they then are taking that LLM and then doing additional configuration so that from a business process perspective, it is. It is able to leverage the data in a way that our, you know, our business, our bionic professionals would do themselves, you know, and that's the key thing where, you know, you get the executive team and they go, oh, chatgpt or Copilot or Gemini. Oh, gee, is it copy and paste? We just put an LLM and then we put it up against our clinical data and we'll get all these great results. And the answer is not quite that way. You know, it's just like if you do your Google search or Bing search or whatever you search with AI, it can make mistakes.

Bill Wallace [00:36:57]:
You need to take the business specific use cases, how the business interpolates the data among the different data sets and sources, and put all that and configure that that is the true training of a model. And the more that you can train a model for a specific business use case, what we've seen is the better the results tend to be interesting. That's cool.

Steve Swan [00:37:23]:
I like that.

Bill Wallace [00:37:24]:
That's nice.

Steve Swan [00:37:25]:
Okay. All right, good. Well, so, I mean, we've gone through quite a bit here, you know, and we're getting. We're getting close to the end. So is there anything more that you think that we should talk about that we haven't covered that our audience might want to hear about? That's either changed over the last year and a half or that's new. That's, you know, helping us or helping you do what you do best.

Bill Wallace [00:37:52]:
Well, I'm proud of the way that you know myself. But more importantly, the technology team that I lead really collaborate with business stakeholders and business colleagues, and we really have a very collaborative mindset. And of course, our business appreciates that. And one of the things that I will say is that AI is changing so rapidly. I mean, even in that example I gave, where you have the engine and you configure it and you get it up and running the engines themselves are being updated so quickly that there is a lot that's evolving because AI is doing some great stuff. And we've already been able to have some solutions, some business solutions that actually have positive roi, which is fabulous. And by the way, that was without having the roi, did not include any headcount reductions. I'll just say that not cool.

Bill Wallace [00:38:58]:
But by the same token, we've already had to go through two additional updates because as the models get updated now we, we bring those newer model, those newer models in and we update with the configuration and now they, and now they improve even more on something that was already giving business value. And so that, and at the same time, not a surprise, all, all these different consultants are reaching out to the business areas and the PowerPoint slides all look great, but you don't want AI to be thousands based on everybody that's ever talked to anyone on the business or for that matter, the technology side. And so that collaboration where you work together and it's not just one solution for everything that's not going to be efficient or effective. But by the same token, you do need to have a reasonably common platform. We have more than one, we have more than one agentic AI solution and we have more than one LLM supporting those. But we've probably had 20 or 30 get proposed to us and so far we've been able to find three that can be used for a multitude of business processes. You don't have to have 20 and, and our overall data platform and how we kind of constrain the data for the AI for the specific business use case has, has also been essential.

Steve Swan [00:40:36]:
Very cool. Well, good, good, good, good. I'm glad. It sounds like things are, things are evolving, things are changing. That's why we did a, you know, a rerun here and had you on for, for, you know, another visit and I appreciate your time and I thank you very much as.

Bill Wallace [00:40:51]:
Thank you before.

Steve Swan [00:40:53]:
Right. I asked one personal question at the end last time. It was about music and if I remember correctly, I didn't go back and look, but I think you pointed at Bruce Springsteen and that was your favorite live show and so on and so forth. This time I'm going to ask you about a movie, right? What movie would you say is one of your favorites and, or your favorite and why?

Bill Wallace [00:41:17]:
Well, okay, so let's see how it works, right? So you know, it is, it is, it is tough to find one movie favorite. But although I go by Bill Wallace, of course, my name is William Wallace and if I had to pick one movie overall, I'm gonna go with Braveheart.

Steve Swan [00:41:40]:
There you go.

Bill Wallace [00:41:41]:
Good. Because Braveheart basically has given 20 years of people I've met for the first time. Oh, do you yell freedom every so often?

Steve Swan [00:41:53]:
Right?

Bill Wallace [00:41:53]:
Yeah, exactly.

Steve Swan [00:41:54]:
That's funny.

Bill Wallace [00:41:55]:
That's great.

Steve Swan [00:41:55]:
I often wonder that. And you know, Bill, never once have I brought that up to you. Never once have I asked. Never once. Because I'm like, he hears this all day long. But I'm glad you proactively brought up so I don't ever have to ask you that again.

Bill Wallace [00:42:06]:
Yeah, well, my, my, my response. My response for many years has been, yeah, I really love Braveheart and my wife wishes I look like Mel Gibson too. So what can I say?

Steve Swan [00:42:19]:
That's good. I like it. Yeah, I like it. Cool. Well, good. Well, I appreciate it, Bill. Thank you very much. This has been awesome.

Steve Swan [00:42:28]:
We'll do it again in another year and a half. How's that?

Bill Wallace [00:42:30]:
Sounds fabulous. Looking forward to it already.