Reinventing Pharma Commercialization with AI with Faruk Capan

Russ Fordyce 🏆 (00:00.997) Hey everybody, welcome back to the winner's circle. I'm here today with Farooq from Eversana. He is the CIO and they have just won one yet another AI excellence award. And he's going to walk us through how to really implement AI in the enterprise. So Farooq, congratulations and welcome to the winner's circle. FAruk Capan (00:18.392) Thank you, Ross. Hi, I am Farooq Kaipun, my current title Chief Innovation Officer of Aversana. We are actually a pharmaceutical commercialization organization. And the project that I'm most proud of it is recently we launched our Aversana AI agency model, which is going to completely change the how we work as a company and many other things that we'll talk about today, but I'm super excited to be here, Russ Fordyce 🏆 (00:44.229) Yeah, now you were actually at, I think what everybody agrees is the intersection of technology and healthcare where we think AI is going to produce the most value for us, right? You're talking about our health, you're talking about our lifespan. We actually think AI is going to do the most good there. What initially drew you into this space? FAruk Capan (01:03.886) Just a quick background, I have 30 plus years in pharmaceutical advertising, marketing, technology. I've seen the internet came through. I've seen social media, mobile. I think AI is probably the most accelerated one in my mind based on my past life. to me, I could have done a beer commercial. I could have been a famous hangout TV commercials. I decided to do the health and I really, that was my first job of course from college, but I really love it. At end of the day, we really do good for our patients and doctors. We educate them, we train them. You know, I try not to do those pesky TV commercials for pharma, but other than everything you see online, everywhere else is pretty much our fingerprints are all over it. Russ Fordyce 🏆 (01:50.844) So you guys are responsible for what you call the commercialization of life sciences or the commercialization of pharmaceuticals. What makes that so difficult for you guys and how is AI helping you solve those kind of challenges? FAruk Capan (02:03.79) Sure. As you see healthcare, pharmaceutical, the first thing comes to my mind, do no harm and also compliance. Finance and companies like us, we have to be very careful what we say, how we say it. And what brings compliance is unfortunately a slow progress of trying to make something a campaign or a marketing happen. And our agency version of our EverSano commercialization, we focus on basically building campaigns. educational materials, websites, everything for the clients. But these things, as this current form takes forever, man, I say forever, it takes months to create an idea and bring to the reality, bring to the patient. I'll give you a very quick one. My first job was at a pharma company. I built the first patient portal and I used to get questions from the patients. I give them to our doctors that they are authorized to answer those and I put them back on the website. that process take three to four months. And I always go back to management that our patients will die by the time I finish with this answer in those Q &A. So I start to bypass the project process and try to publish the answers first, but I almost got fired because I will tell any employee in pharma advertising, don't ever do that. But I did at that time. You have to go to compliance and process, make sure what you're putting out there into public is authorized and accepted by the FDA and others. The NetNet, the speed of the AI right now really got me excited further than anything else. But we have to think differently. We just cannot just take AI and give it to our employees and hope that we'll change the world, we'll become 10x company. That's not the case. Russ Fordyce 🏆 (03:45.83) No, mean, that's really kind of personal productivity gains and you can get so much out of that. But you really need to kind of look at the business and I'm finding that in my own business and in other people's businesses I'm hearing is it seems like everybody's kind of taking a pause and saying, let's step back, look at the business process. Where do I have problems in that business process and then go start to attack those? And it's not just using chat GPT or anything like that. You guys are really doing this commercialization and it's drug development, market access, patient support. Where are you seeing the greatest need for AI's implementation in that kind of whole flow? FAruk Capan (04:25.438) I think when the AI hype came up, first thing agencies, marketing was the old first natural, right? All of us think that AI can create content, absolutely correct. AI can create images, videos. And so our group was the first one to go to say, and I told our team members that we're going to have to eat our own lunch first and then so we can eat others dinner later. Meaning that we know the fact that when we do this AI, we have to look at our cost structure, speed. and quality and everything that clients expecting us to be charging less and current agency models really billable hours. So I'm talking about to tell my team they're going to have less hours to build a client. They're going to be much efficient. And to do that, we had to make a big decision. Try to sort of boiling the ocean inside the company was very difficult. It's a 5,000 plus people organization. It's not a small one. We decided to go with a really a startup mode. where we put a team together and say rethink everything you do from scratch. Don't just, like you said, productive the games of doing faster. Think about a new model, completely new model, new workflow. And that's why we started initially working on the agency. And that was the first one. But as you mentioned, it's not just the agency. Every part of our organization, any organization is going to be impacted by AI. But like your point, pick the pain point and try to really rethink. Completely new model not just try to put holes and let's fix this a couple of the pain points first. Russ Fordyce 🏆 (05:56.338) I like that startup mentality because it does give you that fresh perspective, You're kind of, you know, it's like, how would I do it if I had all these other resources? And it allows you to kind of step back and rethink that. And you guys are in such a complex marketplace with data all over the place, regulatory issues. How are you managing all that complexity, even in that kind of startup world with the AI? FAruk Capan (06:23.656) Absolutely. That's why actually we have a model called human in the loop. It's not that it's a very common term, but we put some really very tough goals to our teams. said 80 % of the activity needs to be done by AI. 20 % will be done by humans and that is a very high quality decisions guiding the AI. Let the AI do the actions, but then humans actually do expertise of medical regulatory creative. strategy, those are the people that in the loop humans that we need them still guide our platform. And we also go from, and it still needs to be compliant, right? Everything we do. we actually, current processes, it's a hierarchical also like almost like linear. You have an idea, you put into a research, then you come up with the concepts, then you go design them. then you send them to approval, then you productize them, meaning that you make the HTML, you make the codes, you put it into the market media and get the loop. And this process, I'm talking about four, six months in pharma because of what we have to go through the steps and compliance. Now the new model, said 80 % AI, 20 % human. And also give you another idea, apologize, the number of people touches the project in this pharma marketing might be 50, 60 people. I'm talking about the medical person, HTML person, maybe a person in India, quality people everywhere, account people, project managers, and bunch of different vendors touching this project too. You're not just the only one. Versus we are telling our clients now, instead of 50, 60 people, you'll have a five to 10 people maximum touching this project as a human. But they are really good humans that they know what they're doing basically. They're specialists. Russ Fordyce 🏆 (08:14.811) So you've spent time kind of in this startup mode, looking at kind of what's working, rethinking it. What does that AI powered platform actually look like inside of the organization? FAruk Capan (08:25.582) Yeah, I think what once we build a platform, by the way, we partnered with Google. We'll talk about this a little bit later. It's important. We have large clients and large clients want scale, large clients want security, large clients want compliance. So you just cannot be overnight success as a startup company. So you have to provide all of those values to your clients too and guarantees. And so we actually put the startup mode and actually we told the startup to say, you're not only building the product or platform. You're responsible even selling this thing and starting to go to existing client and new clients. Make sure we educate them. Ross, we had amazing success. This project is only been launched last August. It's about eight, nine months, I think. And we have 165 clients approach potential clients we talked to. We already have a 10, 15 of them already either piloting or working this project with us with the platform. And we are actually slowing us down to make sure that we can actually convert our existing clients into this platform correctly because it still require people coordination and transformation. It's just not an easy switch button to say, here's how you transform to a whole new workflow model. Russ Fordyce 🏆 (09:40.06) Well, and from what I read through your nomination, you're not just automating, you're really orchestrating this whole kind of life cycle flow through. So you've really taken the time to, you know, not just to break it apart and make it better, or make it faster, you're really working to make the whole orchestration layer different, right? So can you explain what that means kind of in practice? FAruk Capan (10:01.55) Sure. So it's not, these are a bunch of simple codes or snippets that prompts that you put into a wrapper. That's not the case. Your point is well taken. It's basically the system does three things very important. One is the initial project strategy, market research, all the analysis of the competitor analysis, all kinds of market analysis, what happens. We analyze all those things in this part of the system. The second part, your point is orchestration. Okay, what's my creative? What's my channels? What's my patient journey? What's my doctor's, you know, the needs? And then what are the campaigns or what are the journeys I need to build? And finally, actually make sure in this process, everything is compliant, by the way, compliance agents checking everything we do at the beginning. So when we get to the compliant portion is really easy breeze for us to pass through that test. And then finally, also, we are even working to be able to, I will say, not to the press a button and be able to go to the channels to say, here is the Facebook ad, here is the banner ad for this, here is the emails and then orchestrating whole thing. Make sure also loop is coming back to us to say what's working instead of waiting for months to say, did this campaign successful versus will be in days and hours to say it's working or not right now. So that's really the whole system is orchestrating the whole marketing from ideation to execution. Actually, that's what it does. And it's very important. to make it end to end because if you really do what the clients right now want to do a point solution, let me do an agent to build banners. Let me build an agent to do one thing. That is really in the long term is not a 10x solution because you are improving little things. But if each agent, if your system improves each other and tell each other what to do next, it's much better solution than just picking up a workflow and fixing one thing in it. Russ Fordyce 🏆 (11:57.018) Yeah. Now you guys are in a very highly regulated environment, healthcare, and it's really not a place you can afford to make a mistake. How did you approach kind of building that trust layer in and walk us through kind of your decision to pick Google and how that partnership works? FAruk Capan (12:15.766) It's very important. think I kind of make that Google brings the brain and will bring the body parts and everything because we kind of use the term we farm a ties AI. obviously Google trusts us from our expertise in the pharmaceutical marketing and advertising commercialization, but Google also brings scalability security. So for instance, we work with the Google engineers together to build a system from day one to make sure everything we are building is safe and secure. use their own security. a checklist and company to check our work because we deal with patient data, we deal with doctor information. So as you can imagine, one wrong information goes out, one wrong message goes out to somebody. We are going to be liable and important. So we build also the compliance into our system from the beginning, as I mentioned earlier. So everything we do, we check our claims and every data first before we even start a project, make sure what we are doing even compliant. And so that there is not like going back and forth with the medical regulatory legal teams. We work with their IT teams also with Google, make sure that we set the platform is a multi-tenant platform. Everything that is going to be for instance, we don't train the client data at all with our own LLM or system. Everything is confined to that client needs and their data together. They only improve their own data. We don't look at their data, try to improve our solution. So. we put a lot of effort to make sure that client can trust the solution as theirs, because at the end, this platform goes into the client area, then they start to work with it as well, not just us as an agency. So it has to be 100 % secure, 100, I mean, I don't know, can say 100, 99.999 % uptime and security and compliance needs to be built in. That's the way it makes pharma marketing difficult as well as a good challenge for us to solve. Russ Fordyce 🏆 (14:14.989) And I know explainability is now a term that we're all into our AI systems. How did you guys figure out what role explainability has in your AI engine? FAruk Capan (14:25.292) If I'm understanding like this not a black box, this client to say you press and put a document and comes up the solution. Actually, we are platform actually, we actually have a very easy way to start the process to say the client, which data, which content, which things do you trust in that is correct. So make sure we don't un-undate our system with bunch of data just for the sake of it. And then we also explain our, how decision process works and actually each client can customize. So it's not a black box actually. They can see as much as transparency we can provide. What's our system prompt? What's the system does at orchestration? How it does? We explain to clients and they can all adapt to their level of comfort, their level of working. An oncology drug might be very different than a weight loss drug, right? I mean, the way we talk, we make decisions. So we can put extra security layers, extra decision process, but the clients are, it's not a black box. They are really owning the logic. actually in our platform themselves, not just us only behind the scenes. Russ Fordyce 🏆 (15:27.889) As we talked about earlier too, you're doing this whole flow through kind of bringing the product to life and then going through into the of the patient engagement and support process as well. What sort of measurable impact are you seeing from this new AI, from the Eversana AI? FAruk Capan (15:47.586) Definitely. From the agency perspective, as I mentioned to you, 80 % is at least double the triple the speed of what we do. What makes me excited, we have wake up in the morning, when we used to go to client startup, do onboarding. This month we're gonna onboard and do this. Right now in one month timeframe, actually Russ, we can do the whole project, but I'm not telling every client yet because they're gonna be expecting, you know, hey, you know, then let's do it in a month. Of course, we do it in a month, but we need the client to adjust to this speed as well. It's going to take some time to get them to that level. And at the end, it is super exciting because 80 % AI, really, that's what we are sticking to our story that we're to do 80 % and 20%. Much less people touching the project. Decision is going to get faster. And we have some data points that we are making like 30 % cost saving from the get-go. Once they adopt our platform, they're changing 30 % cost savings and they can use that money for more media, more, I mean, yes, true. We are watching that very carefully actually because we have some really good ways to do that economically. But what I meant to say, personalization is very difficult within pharma because you can only create three emails template or three emails in a project. Russ Fordyce 🏆 (16:56.145) More AI, more tokens. FAruk Capan (17:14.126) Now, if the client wants to do 50, 101, n equals one, they can do it because we can help the med rec regulatory process to understand how this is going to work. I mean, you might think that personalization Farag is so 2000s. In pharma though, we still have struggled with being able to customize, personalize these things because of the limitation of the compliance issues. today with our platform, we can actually talk to Dr. Russ directly exactly what he's looking for. what he's asking for rather than standard template emails we are sending them today. Russ Fordyce 🏆 (17:48.326) Right. Well, that kind of goes towards the experience, right? The experience for the patients, for the providers. What is some of that initial feedback that you're getting on that experience? FAruk Capan (17:58.91) Absolutely. Like a year ago when I first started to talk about I can have a lookalike doctors, meaning that we have key opening leaders. We call them like expert doctors. We kind of hire them to talk about, educate the society and patients and doctors speak in conferences. We said we're going to do an AI version of this doctors and they said absolutely not. Nobody's going to ever believe you. Nobody's going to ever watch this videos. We come back and ask this key opinion leaders that would you like to have this technology where you don't have to travel for each video shoot. You don't have to go and fix every time, but we'll actually you'll be in control of your lookalike and then we'll give you what needs to be said and you agree or not. You can make a change and then we'll be available video in this format in this area in this conference at 89 90 % of doctors were all in like I love this. If I can if I have to travel if I don't have to do this, I'll do it for you. And also we thought that patients, they're not gonna really like an AI doctor to trustworthy. But again, it's a high percentage of patients are as long as the content is really good and detailed, they're willing to listen and adopt to this new ways. I believe that again, a year ago, maybe the video wasn't as quality as today, but I promise you since then, video is an amazing way that we can make it reliable. But I want to say one more thing that I still believe the human touch. I still believe that a human doctor, human to human conversations in our business, in our life, but we can do a lot of things automate for our patients and doctors. You you want to really wait on a phone call for 10 minutes on a customer service to find out about your insurance information, or you want to instantly help you with your insurance question and eligibility or pre-approval. I can get those things done in seconds. for you as a patient. Yes, there are maybe some older patients might say, still want to talk to a human being, but I can help them and doctors offices, so many new things with our technology platform. It is very exciting, Russ. I know I'm kind of spraying all over it, Russ Fordyce 🏆 (20:04.877) No, I think it's so interesting because it's like, as you mentioned, we're going to have this human element, I hope for some time. But there are definitely wins where AI can come in immediately and really provide patients information, provide them guidance, very customized, very personalized to your point. And so I think there's going to be this blend as we kind of evolve into this new world, right? I could see us having AI doctors a few years from now or FAruk Capan (20:12.782) Dude. Russ Fordyce 🏆 (20:34.161) decade from now. But there is there's so much trust that's got to be built with humanity that, you know, there's definitely big, you short wins that we can get. But I think long term, I like your approach to that 80 20 kind of approach. So you guys are moving kind of from this experimentation startup world and really putting this into the core of your operations. What are you finding? You know, that's kind of surprising you as you as you move into that kind of operational mode. FAruk Capan (20:39.214) Corner. FAruk Capan (21:04.622) Oh, definitely. You know, chief innovation officer title is kind of a little bit fluffy in general. As you remember, you have a bunch of ideas, most of them will ever make to see the real world. I'm very lucky because our CEO, our management, they all believe in this model. They're committed to it, even though, as I mentioned, we might lose revenue, potentially we might, you know, run into some interesting areas, but they're all willing to take the risk and say, this is the right thing to do. I think that's very important for audiences out there to learn. You have to get a full support. And also, our management we discussed, don't look at this as just that we're going to cut number of people. This is going to be all about jobs. That's not the case. The truth is, yes, we're going to need less people going forward. But the thing is, we still need amazing people in our organization. So let's go be the first ones and get new business and actually keep our company growing. But definitely correct, not 50 people, maybe 10, 15 people we need totally for a project. But let's train our people also in the meantime so that they can also really be marketable for their own sakes as well. So the things that I learned that is new clients are much easier to convince than existing clients, which is oxymoron to a degree because you expect your clients to really believe you. But. The good news actually everybody's willing to do, everybody's really trying to do something with AI, finding those, you know, those leaders, finding those, you know, kind of innovators in those organizations is important, but then also helping them to get to the hump. Like what do I do next? How do I get my organization to try this model? And how do I do it better? And there's also another thing that lot of companies wants to own building AI themselves. They think this is a very core functionality and they're right. They need to be able to learn. They just cannot go find other SaaS companies and say, give me AI is the solution. Because some of them are thinking, actually, maybe SaaS companies are really vulnerable now. I can build my own SaaS models too. And none of them is 100 % correct, because I think all of us, if I have to, what I learned from this thing here is this is ever changing so fast. If you're really not playing and adopting or trying these models yourself, FAruk Capan (23:26.518) You're really going to be very behind because it's just not about people think I can ask chat GPT or grok or cloud and I can do it anytime. It's okay. I can catch up with anybody. There is so much learning grass. You mentioned about your own experiences, like idea, wanting building into the real life is very different. learned selling is different, getting the client transformation. How do I transform the client? Because I just cannot do it by myself from my own side. Lots of learning. Russ Fordyce 🏆 (23:54.256) Yeah, yeah, now you mentioned something a minute ago that as your your fluffy role as a chief innovation officer, you get to kind of you get to, you know, throw darts at a dartboard, right? You get to you get to see what sticks right spitball a little bit. How do you prioritize what you build versus what you ignore? FAruk Capan (24:16.302) One thing very different than us, we have a PNL actually, Rass. Our innovation team is profitable. Our innovation team, definitely your point. We are the kind of a sphere of the organization. The client asks, I want to do market research online. How can I do it with AI? Right? So is this a good idea, bad idea? We go do our own, just typical startup. Is this a, what's in the market? Who's out there? Is this feasible? Clients willing to do it? And we basically take a risk on this thing and evaluate just like another startup as a real company. Are we going to able to make money? Are we going to able to really do this? And then take that project and run through the initial POC, like, you know, just a simple model and get a client feedback and then iterate. So our difference is we are accountable for our own actions at our own department. And we have to be throwing those darts very carefully. Russ Fordyce 🏆 (25:15.343) Yeah, it's just like I struggle personally with I come up with great ideas. I do so say myself, you know, for how to improve the business or how to move, you know, move something along, improve it. And there's there's just so many opportunities. It's it's almost hard to prioritize. And that's where I think your startup approach is really interesting, because instead of just, you know, going after, you know, little little projects that you mount. might mount up to something large, you really took that thing and broke it all down and decided let's re-architect it. So if we have other kind of medium and small chief innovation officers that also have the title of founder and product manager and marketing manager and public relations professional and operations manager and everything else, what advice would you give to them to really kind of start to develop their plan and to really kind of rethink how they're attacking AI? FAruk Capan (26:11.31) Uh, I just want to make sure that we didn't do AI for the sake of technology or for innovation. We definitely see what's coming. So if you, mean, you need to be convinced about your convictions and where the market needs to go, or it's going to go. That's one. Second, it's a business. It's not just let's do this. Uh, you had to put, you have to take an effort to put a business plan together. As I mentioned, we need to have a profit and loss. How much I need to invest in this thing? How long this will take me there. And also. Pilots are very dangerous word for innovation officer because pilots will never, if you call it pilot, it will never go beyond because you're really restricting yourself to one small pilot. But in real life, it might be very different and also to get support. Ours was, as I mentioned, our senior leadership bought into this vision first. Like they really know that AI is going to influence our businesses, our service businesses and they're willing to take the risk. So we have to go sell it up and sell it down as well too. and having that PNL profit and loss approach to your business, it's very healthy because at the end, it's very difficult to convince that if you're committing yourself to say, I'm going to able to influence this business by X percent and grow this business double, I'm going to increase the margin, very hard to go against this innovation. Otherwise, all of us are scared about change. All of us are in nature. that worry about these changes, especially the bigger the organization, harder your job is. If you can convince a small leadership to say, we do this as a startup? Can I be free to promote this, this, and be accountable for it? That makes it very different. I don't think it's going to apply to every model, by the way, because there are sometimes all you need, maybe not this approach. You might want to say this is like a, people call it small I, big I invention in, or, know. You can also say, oh, we're going to do a simple project and still make a big impact in the organization. But for AI type of projects, would highly challenge the companies to think completely of new business model. We have to think that way. That is very scary, but that is, in my mind, most important thing to Russ Fordyce 🏆 (28:29.617) So now that you have this kind of new model and you're working to scale it out to your other customers, what are you looking at next? How do you top that? FAruk Capan (28:37.773) Yes. FAruk Capan (28:43.648) ideas are as you know, for innovation, ideas are great. No, first thing first, we are implementing it, make it real. We have a budget. We have number of clients that we need to convert this year. Number of new clients to attain. And also by the way, I'll just say something very interesting. The new software development is very different with AI. It's not about this all software development life cycle. I'm going to first get the idea, put them into the queue. I'm going to wait. I'm going to go build them. And every two months I'm going to check and every... two months I'm going to make a release. Currently we are making releases weekly and sometimes daily. I mean, that's what another whole thinking that it needs to be changed because people approach with the old mentality of software development. The way our software is changing weekly because the clients tell us something or we see something is to change. We go back and to say, hey, this weekend, let's go make this change because it's important for our clients, that existing clients that I want them to have an amazing experience. And it's silly to prioritize because the thing is ideas are galore, right? I want to make it on AI market research. I want to make AI sales rep. I want to make AI this. We have to be very careful that what is feasible, what is realistic, and also what it makes the most important impact is existing clients on the system. So we are kind of prioritized on, but we are much faster than any product team out there in my mind. And most of the AI companies are like us, native ones. but you have to really act differently. Speed is our only friend because you think that somebody has a good idea, you might catch up with them a year later. I think in a few months, anybody can catch up with you. Anybody can catch up with you. So you have to execute, you have to get enough customers and make them really happy so they don't have to switch. And those are all part of my job. That is not just innovating, putting in the market, but I want to make sure it is successful and constantly iterating, getting better. Russ Fordyce 🏆 (30:39.355) Well Farouk, this has been a fascinating conversation. Congratulations again on all the success and all the success you're having with the platform. I love the approach. We couldn't be more thrilled and we're gonna keep an eye on Eversana like we have been, make sure you guys are staying on course and doing some more awesome things out there. So thank you so much for being on the Winner Circle today. FAruk Capan (30:58.658) Thank you so much for us. Thanks for time. Appreciate it.