Harsh Thakkar (00:01.186)
All right, so in this episode, we're gonna be talking all about fighting medical misinformation and obviously, how can AI help with that, right? So I have no clue about this topic, but I know somebody who knows a ton about this. He was working at a lot of good biotech companies previously in his career as a global and US head of medical information.

And he then started this company called Aingens. And what they're doing is basically having a, they have an AI platform that helps you with content creation, automation, and basically solving this problem because according to a 2022 Harvard study, approximately 76 % of Americans reported seeing health misinformation online, right? So there's obviously a big use case, a lot of use cases there to solve that problem. So.

Without wasting further time, let's go in and have a chat with Ome Ogbru, founder and CEO of Aingens. Welcome to the show, Ome.

Ome Ogbru (01:04.807)
Thank you, Harsh. I'm really excited to be on your show.

Harsh Thakkar (01:08.918)
Yeah, so for listeners and viewers who have never heard about you or never heard about Aingens, what's your elevator pitch about what you're doing?

Ome Ogbru (01:19.059)
So AINGENS stands for AI Narrative Generation and Engagement Solutions. And our idea really is that we can take this incredible technology called AI and attach it to other tools that are needed in the life sciences workspace to really accelerate and revolutionize.

Harsh Thakkar (01:43.341)
Hmm.

Ome Ogbru (01:46.733)
workflows in the life sciences, in the life sciences industry. So as a company, we're not just a one solution company, but rather we're looking for problems that we can solve. And there are many, and that we can use AI to help solve and really change how people work and help them work more efficiently and help them create better outputs as well using this incredible technology.

Harsh Thakkar (01:55.918)
Hmm.

Harsh Thakkar (02:16.654)
So, walk me through, you you started this company about a year ago. I'm not sure exactly the date, but before that you were working in the industry in different roles. So I'm sure you've this idea months or weeks or years before you started the company. What was that journey like? What was going on? Like, what were you researching?

to finally take that step and say, you know what? Now I'm going to go start my own venture.

Ome Ogbru (02:49.501)
Yeah, so for all my 20 plus years in life sciences, we had many challenges. And the technology just wasn't there to help us solve those problems. So for years, I've looked into machine learning and lateral language processing to help with analytics, for instance, and to get those crucial insights out of the information that we receive.

Harsh Thakkar (02:56.43)
Hmm.

Ome Ogbru (03:16.573)
from customers, but the technology just was not at the right spot to be able to do that. And the pain point of creating content, it's a big one because we do it every day, no matter how good your research is, translating that into information that people can actually use, digest, and use to treat their patients and adopt. Therapies is very crucial. It takes time. It takes expertise.

Harsh Thakkar (03:23.885)
Hmm.

Ome Ogbru (03:46.245)
It's painful. It's complicated with all the regulations that you have to go through in live sciences. So the minute I saw AI, I'm like, wait a minute. I think this is what I've been looking for all these years. This is the solution that can actually help us solve a lot of the challenges that we have. And if you look at it as well, content creation is resource intensive. You need people. You need time.

Harsh Thakkar (03:49.25)
Mm-hmm.

Harsh Thakkar (03:59.896)
Yeah.

Harsh Thakkar (04:11.63)
Mm.

Ome Ogbru (04:15.827)
And you need specialized training to be able to do it. But with AI, I think that some of the areas that we've not addressed in life sciences can be addressed. We can create personalized content and targeted content now. So there's really many possibilities that this technology allows us to be able to do. So when I was at my last company, we had a downsizing.

And my role was impacted. as I looked at, okay, what do I want to do next? And what really makes me happy and engaged? It turned out that what that is, is really to build my own company. And so I have a natural affinity for technology. have a natural affinity to solve problems. So I'm naturally entrepreneurial. I thought, well, I don't have to wait till retirement. Maybe this is the right time to.

Harsh Thakkar (04:44.038)
Hmm.

Harsh Thakkar (04:55.182)
Hmm.

Harsh Thakkar (05:06.306)
Mm-hmm.

Harsh Thakkar (05:10.03)
Yeah.

Ome Ogbru (05:12.531)
to really launch my own venture. So that's how it came about. All the problems that I have been facing while in life science, some of my colleagues as well. And now we have a solution. And now how do we take that solution and make it fit for purpose to solve these problems? And that's how Aingens came about.

Harsh Thakkar (05:21.614)
Mm-hmm.

Harsh Thakkar (05:30.602)
Yeah, no, that's a really fascinating story. And I have actually had this debate with, I wanna touch on a couple of things that you said there, right? First, I felt the same way, being in life sciences, in quality assurance, systems validation, doing software audits, infrastructure, qualification. So I was doing a lot of the regulatory.

landscape of how to implement technologies and life sciences in different departments, clean ops, manufacturing, supply chain, all types of systems. So when SAS systems, cloud systems came around, everybody was pushing the narrative. you don't need manual documents because here's a cloud system. You know, it's cheaper. You can collaborate. Look at these reports like it's.

It will do most of the validation. You just have to use the system. So that was a big narrative then and lot of the systems, the cloud systems came into all of these different functional areas. And then now you see AI. With AI, there is this narrative that, you know, it's going to automate everything. But like, you cannot automate a bad process or a bad data set, right? Because it's not what you want to automate. Right? So that's...

Ome Ogbru (06:44.839)
Thank

Harsh Thakkar (06:50.094)
That's one point I want to get into. And the other one which you said is you said something really interesting, which is that you think that you are naturally entrepreneurial. And this is a debate I've had with a lot of people, some close friends of mine who are like in real estate and e-commerce and they're like, oh yeah, you know, I've been like doing stuff since 10 and 11. I know some of these things. So that's the debate I've had with many people. Can anyone be an entrepreneur?

Or can you be a born entrepreneur? So maybe I can talk about that if we have time towards the end. But I want to go into this AI thing. Do you agree that automation is like this push, but there is a lot of other things that we're missing about AI? And what are those things that we should be looking at for a holistic view of what this technology can do for us?

Ome Ogbru (07:46.749)
So that's a good point that you make and automation is definitely part of it, but there's more to it than just that. It's not really that you just automate processes. It's really that the AI helps you with the time consuming aspects of certain workflows. And you have to identify what those aspects are. In content writing, well, anyone who's written content in a scientific

Harsh Thakkar (07:58.606)
Mm.

Ome Ogbru (08:15.171)
sphere understand that well. I have to search for literature and that could take days, hours definitely depending on what you're looking for. And then now I have to find the right information and read it, digest it. And then I have to write my content. That is also painful as well. And through the content writing process, I have to know well, put my citations in there. That is painful.

Harsh Thakkar (08:31.246)
Mm.

Ome Ogbru (08:42.821)
I've always hated doing that as well, all the different types of citation formats. That's painful. I have to edit my content also and also share it with other collaborators as well so that they can, you know, provide me feedback. So when you look throughout the whole workflow for content creation, you can see opportunities where AI can help. It's not really just automating it. It's really having that human in the loop that can tell AI, this is what I want.

Harsh Thakkar (08:42.946)
Hmm.

Harsh Thakkar (08:54.232)
Hmm.

Harsh Thakkar (09:05.249)
Interesting.

Harsh Thakkar (09:10.19)
Mm.

Ome Ogbru (09:12.403)
and this is what I want you to use to help me create that, or this is what I'm searching for, this is where I want you to search for it. And then I will look through that information, pick what works best for what I wanna do, and then give the AI the instructions of what to create. One common misconception is that all you need is chat GPT.

Harsh Thakkar (09:29.219)
Hmm.

Yeah.

Harsh Thakkar (09:36.76)
Mmm.

Ome Ogbru (09:37.963)
or any large language model and it will revolutionize your content creation workflow. And all you have to do is just put in a simple prompt or command and it will create things exactly the way you want it. Well, that is not really true. And it's where we all started. I started there as well. I went into chat GPT, write me a med info letter and it spat out something that like, my God, this is not what I'm looking for. Right? So I'm like, it doesn't work.

Harsh Thakkar (10:03.214)
Yep.

Ome Ogbru (10:06.183)
This technology doesn't work. It's not solved my problem. But in reality, I needed to learn how to use the tool and how it works. And what we really need to really revolutionize content creation in life sciences is fit for purpose tools. So it's the AI plus other components that it uses to create what we really want to create. So that's why we built MacG.

Harsh Thakkar (10:23.779)
Hmm.

Ome Ogbru (10:34.291)
So, MAGES stands for Medical Affairs Contents Generator, and it's purpose-built for people in life sciences to create the content that they want to create. It's a complete AI medical writing and research assistance. It has an advanced editor. It has PubMed and Word Search. It has pre-built scientific prompts, a reference manager, a citation generator, and it has a chart tool as well.

Harsh Thakkar (10:39.383)
Okay.

Harsh Thakkar (10:45.228)
Hmm. So.

Ome Ogbru (11:04.241)
And of course it has Ask Maggi, which is our large language model agent, which uses all these tools to help professionals create what they want to create.

Harsh Thakkar (11:09.112)
Mm-hmm.

Harsh Thakkar (11:16.866)
That's interesting. then you mentioned one thing that was really important there, which is looking at not just automation as an end-all, be-all solution, but looking at the entire workflow and also understanding that every company's workflow for doing the same stuff could be different. So if you look at

for example, cell and gene therapy, the manufacturing and supply chain workflow for a cell and gene therapy company is a little bit more risky, more different, you know, because you're dealing with personalized prod medicine and compared to a generic pharmaceutical company or a medical device company, right? So like it's still manufacturing and supply chain, but you have to like understand, it's like thinking from first principles, right? Like,

yes, you wanna build a car, but you need to understand what goes into building a car, and then how can you sort of understand those each components and say, okay, this component cost a lot and it's increasing the value of my car, can I substitute this or can I do something else, right? So like, it's very interesting that you said that because a lot of people are not doing that. They're just looking for instant solution and just like plugging AI and they don't wanna do that.

work of understanding their process or how their data is going through that process. And that's why the adoption is a bit slower. And I think that's the right way to do it. So I'm glad you mentioned that. I wanted to ask you about MacG. So what are the first few customers that you worked with, what were the good use cases that people were using this product for?

Ome Ogbru (13:05.938)
Yeah, so we have thousands of subscribers now because it has a web subscription option as well, as well as an enterprise version as well for companies who want it more customized to their customer and workflows and in their own IT environment. And what we're seeing that people are really gravitating to the PubMed search. So because it really, what took hours can take seconds.

Harsh Thakkar (13:10.19)
Hmm.

Harsh Thakkar (13:30.296)
Pub mid-search,

Harsh Thakkar (13:35.15)
Hmm.

Ome Ogbru (13:35.387)
or minutes in order for you to find what you're looking for. We also see that people really like the editor as well, because within the same platform, you can interact with the AI and with a full editor to really shape and craft your content with the assistance of the AI as well. We also see that people are using it to create all kinds of things. But one of the most common things that they do is to summarize.

scientific literature. So they find the publication that they want and then they use the AI to summarize that information because that's the basis of any content that you want.

Harsh Thakkar (14:08.129)
Hmm.

Ome Ogbru (14:18.739)
And by summarizing multiple of these articles, you can use AI to now stitch it together into the type of content that you want to create.

Harsh Thakkar (14:18.966)
Right. Right.

Harsh Thakkar (14:26.306)
Hmm.

Harsh Thakkar (14:31.126)
Interesting. also another thing is that when, you know, again, I don't work in this field, but I'm just thinking of the use case that you explained and how somebody would be going through that ideation or content creation. You also mentioned earlier on that in the past, before AI, you would have to list all the citations. You would have to list the references. You would have to look up what that researcher

who study or reading, what other work have they done? Like, do they have authority on that topic that they're talking about to give credibility to their research or results, right? So that's where the question comes about, know, fact checking. How do we know that AI is giving you the right stuff, right? So that's where the human in the loop comes into play. So have you seen when you were...

earlier days when you were building the MVP or stabilizing the product, did you have to train the AI to fact check before it just starts giving out responses? Or did the users have to accept that as something that AI is still not good at and they need to do that on their own?

Ome Ogbru (15:48.851)
That's a very good question. so because of the earlier days of chat GPT and people having to use the training knowledge or repository of knowledge in chat GPT, which often caused hallucinations, people still sometimes believe that AI is not very good. Well, the truth is AI is actually very, very good. But.

Harsh Thakkar (15:55.842)
Hmm.

Harsh Thakkar (16:05.486)
Hmm.

Harsh Thakkar (16:14.99)
Mm-hmm.

Ome Ogbru (16:15.739)
under the right circumstances. And those circumstances is pointing it to the right source of knowledge. So for instance, instead of trying to use the model's own knowledge to create scientific content, you go to a PubMed for instance, and you use that as your source of knowledge. You upload a PDF or any other type of content, and you use that as a source to create the knowledge. Or use your curated websites or web pages that you already know are authoritative and are correct.

and you tell the AI to use that information to create the content that you want. Under those circumstances, AI is actually very good and the hallucination rates are way lower. And then of course, you need human in the loop to review that. And so I think this also touches on a popular misconception, which is, all right, I just tell it, create a document for me and it just creates it. It takes iteration and conversation.

Harsh Thakkar (16:58.414)
Hmm.

Harsh Thakkar (17:11.182)
Yeah

Mm-hmm.

Ome Ogbru (17:15.603)
So you put in a prompt, it creates something, you look for it, it's not 100 % there. And then you tell it, add this, remove that. Why not consider this instead? Or maybe I want to change the tone. With those iterations, you can get to something that's 80 to 100 % of what you want. But you have to be patient and have that conversation with the AI. And when you get used to this, it's actually a lot of fun.

Harsh Thakkar (17:28.535)
Hmm.

Harsh Thakkar (17:43.896)
Mm-hmm.

Ome Ogbru (17:43.923)
Right? Because you have this back and forth communication with a machine like you would have with a colleague, actually, if you were coaching them to do something or create something. And the end result is actually better than either one alone. And I think we'll, you if we get into that a little bit later, we'll talk about that some more. So really we built MacG so that people are able to easily bring their own

Harsh Thakkar (18:04.589)
Yeah.

Ome Ogbru (18:12.403)
into the database so that the AI can use that to create what they want to create. And when it comes to citations, it's automatic. As you create your content, it automatically adds the citations and the format that you want.

Harsh Thakkar (18:14.466)
Hmm.

Harsh Thakkar (18:22.178)
Yeah.

Harsh Thakkar (18:26.958)
Okay. Okay. Yeah. So it's not like, you know, chat GPT where you have to sort of be very clear that you want it to do the search and list the references here because it's scientific content. It automatically gives you the output with the references of how it got the output is what you're saying. Okay. All right.

Ome Ogbru (18:46.413)
Exactly. It's automatically there. It's an AMA format and pretty soon we'll be adding APA and MLA formats and you can easily switch between those two formats as well. And because it's a full citation generator, you can even manually add citations as well and add them in text also very seamlessly and it'll create your bibliography for you also.

Harsh Thakkar (18:51.278)
Hmm.

Harsh Thakkar (18:55.564)
Great.

Harsh Thakkar (19:05.73)
Hmm, okay.

Harsh Thakkar (19:12.354)
Hmm. Yeah, it's interesting because some of the work that we do with companies on computer systems or infrastructure or analytic tools, we've also explored using AI. And I want to hear your take on this because I always get this argument. And the argument is if I...

take data from any existing cloud systems and hook it to an AI data extraction or analysis engine, which is the type of stuff that I do, on a use case by use case basis, right? So we're not a software company, we're a consulting company at Cultivate. And we understand these use cases and say, in the marketplace of AI tools for life sciences, there is this product that could solve your problem. And then we do the documentation and validation if needed.

for the client to implement that tool. And most of the times the clients will ask us, what's the accuracy, right? So like we've got some tools that are 93, 95, 97-ish is the max that we've scratched, right? And this argument is, well, it's not good. If 97 % is not good, I would rather have a human. But like my counter argument is you have FDA,

violations and data integrity violations that say that your humans are not trained on your SOPs and that you have tons of deviation with the root cause human error when you haven't really looked at the real root cause, right? Like I could go on and on. The point I'm trying to make is a human being who hasn't had enough coffee or tea could also slip up, right? So I feel like it's unfair to...

Ome Ogbru (20:39.539)
You

Harsh Thakkar (20:58.412)
What do you feel like? Is it unfair to use that argument against AI to say, I can't use it because it's not 100 % and we're live sciences, so we need everything to be perfect.

Ome Ogbru (21:08.691)
You're absolutely right. So as human beings, we believe that we're the best at everything and a machine can now be better than us. I don't think that's the right way to look at it. The right way to look at it is how can the machine help me be better? That's the right way to look at it. And that's what AI is for. And that's why you need human in the loop. So the combination of human plus machine, the machine helps us.

Harsh Thakkar (21:14.402)
Mm-mm.

Harsh Thakkar (21:25.347)
Mm.

Ome Ogbru (21:39.591)
in bridging some of the mistakes that we may make, especially when it comes to looking at data. Because a human being cannot look at data as deeply as a machine can or as quickly when you're looking at large data sets. So if we use the machine to do the first pass, give us its output, and then we review that output to see what's missing. Most likely,

Harsh Thakkar (21:43.148)
Hmm.

Harsh Thakkar (21:59.758)
Mm-hmm.

Ome Ogbru (22:06.673)
And I think there's enough data out there that shows it already. Cause if you look at, you know, the healthcare space, for instance, where there are now studies that show that AI is actually more accurate than physicians in certain things. But when you put both together, it's actually much better than either one alone. I think that's the right approach, but we want it. And I have seen this, right? People have told me, it did not write it exactly the way I want to write it. So, you know, it's useless.

Harsh Thakkar (22:18.082)
Mm-hmm.

Harsh Thakkar (22:36.174)
Yeah.

Ome Ogbru (22:36.701)
I have to bring them back to? Well, you can't expect it to know exactly what you know. It's not you, right? But if you work with it and give it the right direction and the right information, it can help you with those time consuming steps of what you need to do. And then your job is now to fine tune it and shape it. So it's exactly what you want it to be. So if it gets you to 90%,

Harsh Thakkar (22:44.472)
Hmm. Right.

Harsh Thakkar (23:00.664)
Mm-hmm.

Ome Ogbru (23:05.971)
then you only have 10 % to do to get to 100 % versus you doing 100 % all by yourself.

Harsh Thakkar (23:09.966)
Hmm.

Harsh Thakkar (23:13.87)
Right, right. No, that's actually, I'm glad you ended your statement with that phrase because I read that somewhere. It was probably like a book or something on entrepreneurs and like delegating and using tools and virtual assistants and all of that. And there was a phrase in there which is like 80 % done by somebody else is better than 100 % done by you.

Right, because your time is limited and we're seeing this with a lot of companies and the way companies are operating you mentioned early on on in this episode that you ventured into this journey because of downsizing in your previous company, right? So like how did that downsizing occur? Because we're seeing a lot of downsizing for various reasons. Like if you if your drug didn't pass the regulatory milestone, you lost funding, what have you, right? But I feel like

the future of the biotech companies are going to be looking at doing more with less, right? In the grand scheme of things. And when you think like that, you are naturally gonna have to incorporate AI. I mean, I don't see how you can do more with less without using AI. So I'm very excited to see the new startups that are going to adopt this way earlier than some of the other companies.

who are now trying to catch this train, but they also have so much data, so many systems, so many people who they have to prep before they can say, let's get on this train, right? Like, so that's not an easy conversation to have at a size of company that has 5,000, 10,000 employees, but much easier to have if you have 25 or 50 or less than a hundred employees. Yeah, oh one. Yeah.

Ome Ogbru (25:06.498)
one.

Harsh Thakkar (25:12.224)
Yeah, so I wanna go into a topic that I have been very excited about, but I haven't gone down the rabbit hole, because I just don't have the time and it is AI agents. So what is so great about AI agents? What is this concept? Can you break it down for our users who haven't heard this much or not familiar with it?

Ome Ogbru (25:35.739)
Yeah. So AI agents or agentic AI is the term that people are using. And it really means that it's an AI platform that can really help you automate work. So it can reason, it can work independently. It can complete tasks with little input or no input from the human at all. Now, when you have that,

Harsh Thakkar (25:51.629)
Hmm.

Harsh Thakkar (25:59.566)
Mm.

Ome Ogbru (26:05.147)
it not really helps you to really accomplish certain things. And in life sciences, there are many different applications of how you can use those agents. So let's look at safety. For instance, safety teams have to search the literature, search the web, search different sources to see if there are any adverse events about their product. Go ahead.

Harsh Thakkar (26:12.846)
Hmm.

Harsh Thakkar (26:28.972)
We can actually, sorry, was gonna say, sorry, I was gonna say if you wanna redo that part, I think the phone, was it your phone? Yeah, no worries. You can start from the part like, let's look at a real world example safety and just go from there. That would be good.

Ome Ogbru (26:37.917)
Yeah, I turned it off, but I don't know why.

Ome Ogbru (26:48.603)
Yeah. So let's look at a real world example like safety. For instance, they spend, well, a lot of money, right? Because usually they hire an agency to do this. Finding and analyzing adverse events from the literature or many other sources. You could build an agent that does that and it automatically every day can search all these different searches, all these different sources, looking for adverse events about your product.

Harsh Thakkar (27:14.253)
Hmm.

Ome Ogbru (27:16.915)
and creating a summary and an analysis of what it finds. And all the human has to do is now review that to make sure that it's accurate, to make sure that the output is complete. Another use case for that as well is in medical affairs. Literature surveillance. There's so much literature that comes out about any therapeutic area. You can use an AI to actually do that search for you automatically.

Harsh Thakkar (27:21.198)
Hmm.

Ome Ogbru (27:47.231)
and summarize the information, including sentiment analysis and even recommendations, and use that to create whatever content you want to create or educate yourself about what's happening in your therapeutic area. So there are many, many uses for AI agents. Now, what's interesting is that in MACG, we actually have an agent in MACG. It's not automated, but it's

Harsh Thakkar (28:12.224)
Okay. Mm-hmm.

Ome Ogbru (28:15.953)
That agent is what actually calls on all the other tools, like the citation generator, the PubMed search, and all the other tools that we have inside MacG.

Harsh Thakkar (28:20.268)
Hmm.

Ome Ogbru (28:28.911)
And it does that automatically based on the work that we want it to do, the instructions that we want to give it. So it now knows, OK, I need to search PubMed. So it calls on PubMed, and it uses that tool to do what it needs to do. And it knows that, well, I need to add citations to this. It calls on the citation generator, and all that is generated for you. But the next step is making

Harsh Thakkar (28:54.67)
Mm-hmm.

Ome Ogbru (28:57.759)
using it as an autonomous agent. And that is coming down the road. So as the large language models improve, we will also build in automated features into MacG as well. So we're close.

Harsh Thakkar (29:02.574)
Hmm.

Harsh Thakkar (29:14.39)
Yep. And another, know, those are some really great examples. Another one that I am really excited about hearing when I hear about AI agents is monitoring, you know, regulations and guidances for different worldwide authorities, right? So like, it could be for a specific therapeutic area, or it could be for...

specific technology, like, if you want to understand what the FDA is putting out about AI or what is EU putting out about AI, did they put a guidance document or did they launch some act or are they having some industry discussions? I know lot of this can be done by other methods like Google Alerts or like subscribing to different feeds and all that stuff, but

to be true, those are not truly autonomous, right? Like you still have to put in the effort and at least they're not reasoning to a level that it sounds like an AI agent would be able to reason. So yeah, that's another example that I'm really excited about, especially in my field of work, which is heavy around, you know, studying the regulations and what impact does it have on a client to apply those regs.

Ome Ogbru (30:35.251)
You're absolutely right. what's different about using AI, it's not only going to search for the information, it can also categorize it for you, it can analyze it for you, it can summarize it for you, it can even suggest actions for you as well. And it help you with those actions.

Harsh Thakkar (30:41.826)
Hmm.

Harsh Thakkar (30:49.768)
Right, right. And that's the part that's lacking in a Google Alerts or other feeds, right? Because those, know, stuff like Google Alerts or these other news feeds, they are just opening the fire hose to say, here you go, you wanted the information, right? Here you go, go knock yourself out, right? But that's exactly where the agent comes in, because the agent can

say, hold on, you've got like 300 links. You're not, no one in the right state of mind is gonna click through all those 300 links and figure out, let me do that heavy lifting for you. now these are those five that you should really look at, right? So, so yeah.

Harsh Thakkar (31:39.95)
Yeah, it's been great learning about your product and a Mac G and what you're doing. I want to put you on a spot for one second if that's okay. You've talked a lot about AI and especially how it can be used in the scientific research content creation area. When you are working today on making the product better or

Ome Ogbru (31:50.739)
F

Harsh Thakkar (32:06.698)
gathering other use cases from your customers and users and taking that into account. What is one prediction that you have about AI that you think might happen in the next 10 years, but you're not seeing it yet?

Ome Ogbru (32:23.379)
Well, 10 years is a long time. And people say, you know, the next two years we're going to have generalized artificial intelligence. Well, I don't know about that yet.

Harsh Thakkar (32:27.34)
Five, whatever.

Harsh Thakkar (32:37.664)
I, same here, yeah, I don't think that's, I would be scared if that happened, but I don't, I don't think that's, that's happening, but maybe I'm wrong.

Ome Ogbru (32:42.995)
You

I think where we can move incrementally is bigger adoption of AI. So a lot of companies now in the life sciences space are moving from that curiosity to experimentation. And some are now at the point where they're like, OK, we understand what we want to do. We understand what problems we want to solve. Let's now, you

Harsh Thakkar (33:03.822)
Hmm.

Ome Ogbru (33:16.893)
build a tool or buy a tool that can help us solve those problems. So, and those companies come to a company like mine. How can you help us solve this problem? So I think in the years to come, we will have more adoption. And I think our perception and attitude toward AI will change as well. Because now you still hear a lot of discussion about, know, this can AI really help me? Does it really work?

Harsh Thakkar (33:21.805)
Hmm.

Ome Ogbru (33:45.233)
It has hallucinations, there's risk. But I think as time moves on, it will become the fabric of how we work. So people will be reaching to AI first before they do anything. And if you're not using AI, people are gonna wonder why. AI will become a best practice. And now with that, with it being a best practice,

Harsh Thakkar (33:55.118)
Hmm.

Harsh Thakkar (34:05.147)
Hmm. Right. Right.

Ome Ogbru (34:15.375)
In the content creation space, one of the ways that AI can really make a big impact is the development of automatic, targeted, personalized content for patients and health care providers. And I think in the years to come, that's where we're going to be. I don't think that the current

Harsh Thakkar (34:31.192)
Hmm.

Ome Ogbru (34:41.841)
the current approach of I have to go search Google to find information and then assimilate and digest all that information by myself manually. Or I have to call the pharmaceutical company to get information. That process is long, it's painful. I think where we need to be, and it's possible, we're not that far off really, is anyone who needs to learn about anything.

Harsh Thakkar (34:52.622)
Hmm.

Ome Ogbru (35:12.401)
with AI can get that information.

Harsh Thakkar (35:16.138)
Right. And the reason why I like this future that you've painted is because as human beings, we find more joy and purpose in fulfilling our curiosity or when we think about something like, maybe I should improve this process or maybe is there a better way to write this research article than what I'm doing? We're not designed to do manual work.

Right? We don't like doing it. A lot of the nature of our jobs has that component in it. But even like today, you know, being in, let's say I'm in quality or in regulatory, I'm reviewing documents on my computer and I'm making comments. this doesn't align with this regulation. Do I love doing that? Yes. I mean, it's my job, so I'm doing it. But would I be opposed to using AI tools that are going to do that for me?

and tell me, I scanned this document. Here's the seven areas that are highlighted in yellow that I think don't align with XYZ regulations. You need to comment on those, right? So instead of me reading the whole article. So I feel like it's gonna make us more, it's gonna make us like our jobs a lot more because it's gonna take out all of this manual work that we don't like doing. We're not designed for doing that.

And it's going to bring this fun element of, let's go experiment, find a low risk use case, see what breaks. Then if it works, go to a medium risk use case. Then if it works, go to a high risk, right? And then go through that workflow. Cause every week you're testing something new and you never get bored with your job. So I feel like that's what the promise is for people, those who want to adopt this path.

Ome Ogbru (37:15.415)
One good point that you make is making your work fun. And that is honestly, every time I use AI, I'm happy. It brings a smile to my face, right? Because it's now something that could have taken me three hours has now taken me 10, 15 minutes to achieve. And now I can go do something else, right?

Harsh Thakkar (37:19.116)
Hmm.

Harsh Thakkar (37:26.272)
Yes. Yes.

Harsh Thakkar (37:41.816)
Yeah.

Ome Ogbru (37:42.607)
And I can just imagine that if I was still in medical information, just how much I could have accomplished having a tool like this. I think, and I don't hear enough about it, but I think AI really improves job satisfaction. Right. Cause if you take that grunt work out of it and content creation is grunt work, you know, your fingers are typing, your fingers get tired.

Harsh Thakkar (37:48.078)
Hmm.

Yep. Yep.

Harsh Thakkar (37:56.878)
Hmm.

Ome Ogbru (38:10.931)
It's a long process and if AI can make that more enjoyable and more seamless, I think people will be happier with the things that they do. And then they can even focus on more strategic things, which really helps the company.

Harsh Thakkar (38:25.334)
Yes, absolutely. Yeah. I think that's one of the ideas that I'm excited about in this industry is just because a lot of our roles, I talk to people even outside quality in different functions and some of the stuff that we're still doing is just so outdated. And I find it hard to believe why aren't we adopting more tools that are out there? Yes, I understand there's...

You know, we can't adopt tools like an e-commerce or retail company because we're in life sciences and there is patient and data integrity and all this other stuff that we need to check. But I still feel like, you know, we could do a lot more, right? And I'm excited about companies like Aingens who are, and you who are coming out here and painting this picture and helping people realize that, there is other stuff out there that you can be more creative and...

get some of that joy back in your job. So like you're coming every week and you're experimenting on something new rather than coming in and doing something that you've done eight times before and you're just trying to get it done, right?

Harsh Thakkar (39:43.316)
Yeah. So I want to thank you so much for coming on here today and sharing everything about what you're doing, how you started your journey and insights on this topic about using AI for all the different use cases that you talked about and the product that your team is working on. For somebody that has watched this episode, listened to this episode,

and it's evaluating their current stance on either using AI in a different way than they have been using, or they're not using AI and maybe this is a podcast that they feel like, okay, I should really look into this. What's a big takeaway that you wanna leave the audience with before we drop off?

Ome Ogbru (40:36.775)
Yeah, I think the first thing is just do it. Just try it. There are low risk options out there that you don't have to expend a lot of resources. You can start by using free tools just to see how they work. You can try MacG. You do not have to go through an implementation process. You can simply just subscribe online and try it out and use it and test it.

Harsh Thakkar (40:39.66)
Hmm.

Harsh Thakkar (40:50.926)
Hmm.

Ome Ogbru (41:04.571)
and see if it works for your workflows. And when it does, you can really understand, okay, what else do I need in addition to this to really do what I want to accomplish in my organization? I think the bigger risk is not trying. There's really no downside to trying. And because without trying, you don't really understand what it can do and what problems it can help you solve. And then I think the second thing is really people should educate themselves.

Harsh Thakkar (41:22.038)
Hmm.

Ome Ogbru (41:34.215)
There's a lot of noise out there and there's a lot of misinformation about AI. So understanding what AI can do and cannot do really will help people see how they can adopt it. And then the other thing is it's not going to solve everything for you. It's not meant to replace you. It's really for you to be able to work with it to accomplish what you want to accomplish and do it more.

Harsh Thakkar (41:42.542)
Hmm.

Ome Ogbru (42:04.199)
more efficiently. And then when you're ready to embark on that journey, understand what problems you want to solve. And make sure that the problem you want to solve, you can measure the outcomes, right? Because you want to get your ROI. Make sure it's something that you can measure. don't set an objective of, want to reduce my team size.

Harsh Thakkar (42:04.462)
Hmm.

Harsh Thakkar (42:15.885)
Hmm.

Harsh Thakkar (42:33.474)
Hmm.

Ome Ogbru (42:34.353)
Right? By 50%. That's not the right way to go into this. But rather look at something like, well, it's taking us five days to create a certain piece of content. Can we reduce that to a day? Can we reduce that to a few hours? And would the quality be the same or even better than the way we were doing it before? This is measurable. It's a tangible output.

Harsh Thakkar (42:50.574)
Hmm.

Harsh Thakkar (43:03.374)
Hmm.

Ome Ogbru (43:03.857)
I'll start with those types of use cases, prove that it works in those use cases, and then you can expand out. And very importantly, just because you build it doesn't mean that your teams want to use it. So there are significant change management that has to take place in your organization, because you're changing the way people work, the way people think, and you're teaching them something new, and they need time.

Harsh Thakkar (43:19.694)
Hmm.

Ome Ogbru (43:33.875)
to be able to learn that, adapt to that, accept that, it works. So you have to build that into your planning as well. And of course, train them, give them the training that they need, give them the bandwidth to experiment and discover on their

Harsh Thakkar (43:45.422)
Hmm.

Ome Ogbru (43:53.755)
So all of those change management principles are very, very, important. Because what you don't want is to spend a lot of resources and then people don't use it because they're afraid of it, don't understand it. They're thinking about, well, is this going to replace me? So you have to think about all those things before you expend a lot of capital in AI.

Harsh Thakkar (43:54.114)
Yeah.

Harsh Thakkar (43:59.905)
Absolutely.

Harsh Thakkar (44:04.398)
3

Harsh Thakkar (44:15.918)
Yeah, no, those are some amazing points. I think the big ones for me are, don't be afraid of trying. It's not that you, just because you're trying something, that's your new reality. You can try it. It's like you're going to a restaurant you've never eaten. All you do is you like eating, let's say chicken or turkey. You've never tried fish, but you go to this amazing,

restaurant, you know, by the ocean and you know, everybody in town is talking about how great the salmon is and you just give it a try now. Now you have one extra protein that you can add to your diet, right? so, so I feel like we, you know, as humans, we are sometimes afraid of change because of what it might do to our day to day routine, because we like consistency, we like predictability, we want to know what's happening next in our life, right?

and or we want to know what to expect in our job. But I feel like that's sometimes that can stop you from growing. That's my perspective there. So I feel like, you know, be open to trying it. That's a big takeaway. And the other one is on the training side, as you mentioned, understand what it can do and what it cannot do, because you going in and trying to use it to do something it cannot do and then blaming it that it doesn't do.

That's on you right like it's again. I give this example these random examples you go, you know, you go somewhere and If you don't like let's say you go to I don't know Let's say you go to get a haircut, right and you don't tell your hairstylist what kind of haircut you want They're not gonna know they can't read your mind. You need to tell them hey I need a haircut because I'm I want to look professional or

Hey, I'm going to a sporting event. I need something with my team's logo and then I'm going to, you know, buzz my head, like the hair off. like, looks perfect. Like you have to give them your requirements because otherwise they're not going to know. And I feel like that's the same thing you want to approach AI with is just what understanding what it can and cannot do. Otherwise, like you said, we're just going to end up wasting resources and then not expect the outcome that we want from it. So.

Harsh Thakkar (46:42.55)
Those are really great and I will also add the links for MacG in the show notes of the episode. So anybody who is listening to this, if you guys want to give it a shot, we'll put those links in the show notes. And then where can people reach out to you if they have more questions about the product or they want to talk about some potential use cases of whether it can do those or not?

Ome Ogbru (47:09.799)
Yeah, so you can go to our website, Aingens.com. So A-I-N-G-E-N-S dot com. You can also from there access MacG as well. And you can learn all about it. The URL for MacG is A-I-N-G-E-N-S dot com slash MacG. You can also connect with me on LinkedIn as well. And either through our website or directly, you can send an email to my team at support.

Harsh Thakkar (47:13.016)
Mm-hmm.

Harsh Thakkar (47:29.016)
Mm-hmm.

Ome Ogbru (47:38.575)
support at Aingens.com and we'll be able to support you with your needs.

Harsh Thakkar (47:44.202)
And is your team is mainly in US or are they like in Europe and other parts of the world? We're good. Okay. All right. Yeah. It's been great. Ommi. Thanks for coming onto the show and being so open and sharing everything that you're doing. I'm very excited about this field and I'm rooting for you and your team and the product to see how it evolves and what more it can do. So I'll be...

Ome Ogbru (47:49.191)
We're global. We're very global.

Harsh Thakkar (48:12.238)
I'll be following you guys and looking at all your company developments and everything you're doing. So wishing you the best.

Ome Ogbru (48:18.269)
Thank you, thank you, Harsh. It's been a pleasure having this conversation with you. Love, loved your examples. You made me hungry actually. think I'm gonna go try out that new restaurant.

Harsh Thakkar (48:24.609)
Yeah.

Yeah, yeah, yeah, it's it's you know, you have to you have to make this content a little bit more entertaining, right? So like I've learned using analogies is the best way to get your point across

Ome Ogbru (48:42.949)
Yeah. Now, Harsh, you may want to change the focus from misconception because we, or misinformation because we didn't talk about it.

Harsh Thakkar (48:52.448)
Yeah, that's fine. can look at this and the stuff that we did. So I can stop the recording here.