Aired:
November 27, 2024
Category:
Podcast

Why Data Strategy Must Be Tied to Business Outcomes

In This Episode

In this sharp and strategic episode of the Life Sciences DNA Podcast, powered by Agilisium, the message is simple yet powerful: data efforts mean nothing unless they drive real business outcomes. It's a call to action for biopharma leaders to stop chasing disconnected tech trends and start linking data strategy to meaningful impact.

Episode highlights
  • Encourages teams to move away from “cool projects” and focus instead on solving core business problems that drive results.
  • Reinforces that every data strategy must begin with clarity—what problem are we solving, and for whom.
  • Explores how a connected data approach can improve everything—from R&D efficiency to commercial performance and supply chain reliability.
  • Dives into how to define success in a way that goes beyond dashboards—by tying data initiatives to KPIs that leadership cares about.
  • Urges C-suites and data leaders to co-own the transformation, ensuring data becomes everyone’s responsibility, not just IT’s.

Transcript

Daniel Levine (00:00)

The Life Sciences DNA podcast is sponsored by Agilisium Labs, a collaborative space where Agilisium works with its clients to co-develop and incubate POCs, products, and solutions. To learn how Agilisium Labs can use the power of its generative AI for life sciences analytics, visit them at labs.agilisium.com.

Amar, we've got Kannan Raman on the show today. For audience members who don't know who he is, who is he? Kannan has the North America delivery for professional services in healthcare and life sciences practice at Amazon Web Services or AWS. He has more than 25 years of life science experience and has deep expertise in deploying digital transformation at scale.

He's led large and diverse global teams and worked for Fortune 500 companies in both technology and life sciences, including Digital, Pfizer, Johnson & Johnson, IBM, Cognizant, and Hitachi. And what are you hoping to hear from him today? Kannan has been working at the intersection of technology and life sciences and has seen close up this transformation to tech bio and how data is transforming the life sciences. So I would like to get his thoughts on this transformation, what information is enabling today, the challenges we're facing, and where we may be heading. Before we begin, I want to remind our audience that if they want to keep up on the latest episodes of Life Sciences DNA, they should hit the Subscribe button. If you're enjoying the show, give us a like and let us know your thoughts in the comments section.

With that, let's welcome Kannan to the show. Kannan, thanks for joining us today. Can we talk about the current dynamics in the healthcare industry and specifically how that affects life sciences? Hey, Amar, I'm happy to be talking to you. And when we talk about macro trends in the industry today, and I'm going to call out the common ones, which cuts across all industries, which is inflation and interest rates. Not specific to healthcare, but the situation is the cost of healthcare has gone up. And if you look at what has happened, there is some analysis and data to kind of suggest that in the past 10 years, the cost of care, in fact, insurance premiums have actually gone up by about 50 % in a 10-year timeframe. And that is pretty large and significant.

And if you also look at how it is poised for 2025, there are estimates in terms of 5 to 7%, right? Somewhere between 5 to 7 % is the cost of insurance coverage. This puts a lot of pressure on employers from a health care perspective. So that's one macro trend, right? In terms of secular trends, right, and I'm going to be very specific here about the US healthcare system. It is very complex. are a lot of stakeholders and the interaction and interfaces between those stakeholders is quite significant. What that does is in terms of number of patient touch points and caregiver touch points, that is quite high in terms of the patients managing it through their treatment pathways. Now, there is this whole aspect of value-based care wherein all of us want to ensure that the healthcare ecosystem is able to provide quality care at an affordable cost. But there is more that can be done in terms of scaling it up.

So technology is definitely going to play a role in that. Now to talk about tailwinds, and I'm going to be again here specific to life sciences. And it looks very promising actually. Monoclonal antibodies, as an example. Development of protein in a lab that can actually bind to a disease protein or a disease cell like cancer cell and control that, right? mRNA, another platform that's available today in terms of messenger molecules basically carrying instructions to enable natural mechanisms of the cell to actually generate the protein needed, right? So that's the other mechanism. And when we talk about gene therapy - gene therapy is another promising area that we are seeing today, both from a gene therapy wherein putting a genetic material in the cell to make corrections and gene editing. And recently we've seen CRISPR Cas9 based treatments getting approved. So the way I kind of tend to look at it is from a macro trend perspective, companies are extensively working on putting platforms out there, investing heavily on these platforms. And over a period of time, it is also going to enable us, enable companies rather, pharmaceutical companies, to move towards scale in personalized medicine, which makes this whole thing very exciting. And technology is playing a critical role. Today there is technology available that actually helps along with computational biology, data and artificial intelligence to actually come up with these platforms and enable those platforms ... Yeah, so we're going to talk now about how this ability to manage the large data sets and technology and how that's changing life sciences and the different impact, the different ways it's impacting the industry and how it may be changing what's possible. Right. You just talked about mRNA, cell gene therapy.

These are new modalities that weren't there a few years ago, right? So it's a very exciting time. And we've also on these podcasts, we have talked about some of these new modalities that are coming up. perhaps we can begin with like a high level overview of where do you see the biggest impact from these changes, especially happening in data, AI, technology, on the life sciences? Absolutely. It's a great question, Amar. Now, if I take a moment, right, to actually talk about data strategy, per se, and then we will build this around that. We all have been talking about data strategy for a while now, in terms of data architectures, data management, data governance, all of that. What's different now from a modern data strategy perspective is how can we actually ensure that data strategy is based on business outcomes?

That's a very different way of putting together a data strategy that is directly connected to the business outcomes. So think of this as having multimodal data sets. How can the multimodal data sets actually come together and integrate in such a fashion so that it can actually impact business outcomes? So can you give an example of a business outcome that you're talking about? Absolutely. And this is where I think the whole concept of data as a product and monetizing data comes into play. And I will kind of explain, give you one example of monetizing data, a classic case in drug discovery. Some estimates actually put that if you take the top 10 spend across all pharmaceutical companies in terms of research and development, it is greater than probably more than $120 billion. And some estimates even project it as over $150 billion.

This is an area where a significant impact can be made leveraging technology and data. Now, you look at in this particular situation, if you look at a use case from a drug discovery standpoint, imagine lead optimization that can actually be achieved by leveraging genomics data, the power of cloud, because now you have unlimited amount of computing power at affordable cost as you need it. And then bringing other data sets like molecular data sets and protein properties and all the public research that is available today. Bringing all of that together in an integrated fashion and having objectively reduced the time to patents. Now that is effectively monetizing the data and ensuring that on one hand, getting a product out there in the market definitely helps patients and that's enabled. But it also, reduces the time to patents. Leveraging all this actually also brings in the revenue faster, which is important. Yeah. And when you said like lead optimization, so for audience members who may not know lead optimization, you have a lead candidate, which is going to be a drug molecule. But so you're trying to optimize it because the first one that you find is not the best.

You're trying to make sure that you're changing that so that becomes the best, right? And that's what you're talking about. Like if we can make that faster, we can get the drug out faster. Absolutely. Absolutely. And some classic examples in terms of molecular docking and post-predictions. This is a situation wherein you can  leverage computational biology along with AI ML to determine how a ligand is actually going to dock or orient itself to the disease protein. And with a fair level of probability in terms of the success rate of the right kind of docking. So what this does is you are actually doing a lot of work in silico so that you're actually taking thousands or millions of molecules and basically saying molecular structures and basically saying these are the few molecular structures that hold the most promise in terms of protein binding, which is a very important step in controlling the disease protein. So I think that's one example. There are multiple examples in the discovery space, including search and summarization, which is another example where as discovery begins, there is a ton of information that is out there around a particular therapeutic area or a set of molecules, which we can use search and summarization and generative AI to actually bring in those aspects. And so what you're saying is that data strategy now is not something that's like in its own world, but it is now getting tied to these specific business outcomes. And so it's much more streamlined toward the impact then. That is correct because - and the whole idea being, how do we treat data as a product? Now, there are multiple schools of thought here, at least two schools of thought. And I tend to take the hybrid approach, and I'll explain that in a moment. On one hand, when you look at the pharma value chain, one might argue that each of the parts of the value chain, like discovery, clinical, manufacturing, supply chain, commercial, each one of them have those nuances.

Maybe there is a thought process in there that it needs to be treated with those nuances, right? And the data needs to be owned. The other way of looking at it is, can all the data be brought together in a very centralized manner and managed and governed in a centralized manner? Now, there are companies that actually are doing that, centralizing the data so that you have a central governance structure, a central set of tools and makes life easier in terms of integrating those data sets. But it is not always possible. But what we can do here is we can ensure that even if data may be sitting in certain silos and need to be treated with certain nuances, you could have a common data strategy. You could have a common data governance.

And you can also use a common set of tools to the extent possible and be able to actually interconnect data sets that are needed to be interconnected. Like for example, in the pharma commercial space. When the product is in market, understanding real world evidence, right? But the real world evidence information might also be useful for folks sitting in the clinical space. So there must be an element of data integration that needs to happen.

So I would say that if we are able to centralize data and connect it in one central place, that is good. But that's not the only way. Sometimes we've got to take the hybrid approach of basically saying, even though it may be sitting in various different parts, how can we have a common data governance model, common security model, and have the ability to bring data together on a need basis? And that's going to be critical  to the business goals.

Yeah, because see the way I think about it is that  on the research side, you have a lot of this omics data, right, genomics, proteomics data, which is much different from the clinical trial data, which is much different from the sales data that we get. And that's different from the manufacturing data, right? So  like  it's hard to have like one common data model for these really very different data sets. But as you're saying, the needs in terms of data governance and data quality are the same everywhere. I mean, it needs to be there, right? So I like this idea that you talk about, you know, it's yes, these are different data sets, but then we need to have like the common data governance model that should be coming across. Absolutely, absolutely. You're right on the mark. And that's, I think, critical. And again, I would reemphasize the fact that from a modern data strategy perspective, it has to be grounded on business outcomes.

Okay, so how are you like now see data strategy, right? I mean, you've been in the industry for such a long time. Now in the last, let's say a few years, let's say, you know, two to three years, how like, what are the some of these changes that you're seeing? Are you seeing companies putting data like companies taking this hybrid approach for  data or like, you know, this business, they're tying that like, what are some of the changes that you see these life sciences companies making in the last few years? Absolutely. It's a great question, Amar. Let me also take a moment to explain what I call as a perfect storm for innovation. Now, we do have cloud computing. We have rich datasets. You mentioned about genomics data, multi-omics basically, all different datasets, omics-related datasets.

You have EHR EMR data, and you have so much of publicly available public domain data sets that can be leveraged for various different purposes. So you have that, and then you have the advancement in AI and Gen.AI. So when you combine all of this, I call this a perfect storm for creating innovation. And companies are kind of realizing that. And it requires a very sound data strategy to enable that. So it has to be very intentional, right? In terms of what business outcomes do companies want to achieve in each and every part of the value chain and how does that drive the data strategy of bringing data sets together as needed, right? I mean, like you said, not all data sets can be connected and there may not be a need to connect all the different data sets.

But it is actually understanding that common denominator of data sets that need to be connected and prioritizing that so that the use cases can be built on it and the use cases can ultimately be rolled out for actual use. Okay. Now I see it's easy to connect to business outcome on the commercial side, but when we're talking about on research or development side, how do people come up with this? Okay, well, this is the exact business outcome. What are some of the ideas that you have or what have you seen in the industry people? Yeah, so let me take another use case, right? And again, it is around lead optimization, leveraging Gen.AI now, since that's the technology that is pretty prominent right now. Think about how actually we can bring in Gen.AI across datasets and understand the properties that are required to a molecular structure. It is basically leveraging Gen.AI to generate molecular structures. That is the best chance of binding to a protein. So it is of generating. And once that generation happens, you can then again use in silico to go after molecular datasets and figure out which ones come very close to the one that the Gen.AI has identified. So this is actually reverse engineering and reducing the amount of time for lead optimization and then basically taking the next step. So that it is very intentional in terms of what we're trying to do to enable identifying molecules that is the best promise to go into clinical trials.

And when you're talking about clinical trials, here again, there are many use cases that we can talk about, right? One is the- Yeah. So maybe before we go into clinical trials, right? Let's think about what are some of other use cases? And maybe I can maybe generalize this question a bit more. we talk about data. Now, of course, now there's Gen.AI, which is making a big time entrance in life sciences.

And maybe you can talk about the use cases for like both, you know, data, cloud,  gen AI. But before we go into that, how are you seeing the adoption of gen AI by the life sciences companies because, yeah, they want to do it, but it's not that easy to do it because of a lot of compliance and also it's kind of wild west. So like, what are you seeing right now in the life science industry about the adoption of gen AI and how it's going? Yeah. So there are many use cases, right? I mean, like you mentioned, on the pharma commercial side, search and summarization and content creation is pretty good, right? And it definitely saves a lot of dollars and time to leverage in AI to create content on the pharma commercial side, because there is medical review, there is legal review. All of that can be automated to a large extent.

Right? Yet there might be human intervention required. Until all confidence is gained that a particular model is actually going to take care of it on its own, there is always going to be feedback coming in from intervention. Yeah. And we're not there yet. But are you seeing a lot of these companies jumping into the Gen.AI?

To what extent though, like are they just doing maybe just like some point solutions right now or they are like implementing platforms or so like what are your thoughts there? Yeah, where are we? It starts with point solutions, right? It definitely starts with point solutions, but we're also seeing how those point solutions can actually scale. And also we have the tools today, right? In terms of the travel augmented generation, which is bringing in context sensitive information to the large language model so that it can pass through that and make it and reduce hallucinations, right? So that's one part of it, right? We're also seeing how multi-autonomous agents and multiple autonomous agents coming together in a workflow, leveraging technologies like LandGraph, right? That makes it more, complexity of the use cases can be much more, and it still can actually drive through. So we are starting to see those kind of exciting use cases as well where the latest technology that is available is getting used. We're definitely moving towards productionizing some of these models from a customer standpoint. What I've found very interesting is that the idea of this Gen.AI agents that you talked about. Gen.AI agents, I think that's a concept that came up just a few months ago and it seems to have taken the industry by storm. Now we're talking about agents everywhere, right? Is that what you're saying now? Yeah, yeah, we're seeing that, right? And that is exciting to me because it is able to actually leverage LLMs, but take it to the next level of having a workflow of multiple agents working in tandem. And each of the agent is actually an expert in certain part of the entire...value chain, right? So, yeah, yeah. So, each of them is doing a specific small task that they're experts in, but then we can bring in the agents together and then that way we can get some complex activities done, right? Absolutely, absolutely. And when you start doing complex activities that are kind of human-like, know, it definitely gains traction and, you know, it actually solves for a problem. we are definitely seeing some of that.

Right? Gotcha. So now as we start talking about these use cases, right, and irrespective of data, cloud, technology, Gen.AI, where you're seeing those. Maybe let's start with research. So in research, you talked about the one use case, which is developing these new molecules, right? Which probably is probably the most valuable use case you can think of because we can go with new drugs with that, right? What are some of the other like really kind of impactful use cases that you're seeing in the research area. Yeah. So I did mention about summary and summarization, search and summarization. Yes. Now, you can actually put conversational AI on top of that. Because the amount of efficiency that scientists can actually gain by having some kind of a conversational natural language, kind of a method of parsing through that data, because it's a toll of information coming from various different places. That's a very effective use case in terms of leveraging AI and Gen.AI and conversational AI in particular. We're seeing that. that's some of the, I would say, low-hanging fruits in terms of how we can leverage AI and Gen.AI and data that is available to make life easier for people, right? So that's one aspect of it, yeah. Okay, okay. Moving to the clinical trials, and I know you started talking about that. So moving to clinical trials, what are some of the biggest use cases that you see? Yeah, so there are multiple use cases on the clinical side, right? One example that comes to mind is digital clinical protocol design, right?

So it's creating the clinical protocol design for a new clinical trial. How can one leverage all the previous clinical trial designs that have been done and also results of those trials and bring all that to the leveraging generative AI to create a new design protocol? And of course, it probably does a good percentage of the work, then there can be human intervention to some extent to finalize the protocol design. That's one aspect of it.

As we go about it, we also look at other use cases from a patient recruitment standpoint. You have the EHR, EMR data, genomics becoming available, leveraging all of that and identifying the patient population that can be approached for say a clinical trial in terms of patient identification, the cycle time significantly.

We're also seeing examples of patient assistance, right? Okay. traditional chatbots because I think patient retention is a very important aspect in a clinical trial. It's a cost of actually replacing, due to dropouts, replacing a patient in the middle of a clinical trial. It's four times initial recruitment cost.  So if there is a way in which we can actually do well in terms of retaining patients, and one of the aspects studies show is how easy one can make it for patients to kind of report back, right? So imagine a personal assistant or a conversational AI, when they have to respond to their daily questionnaires, they can do it leveraging a voice-activated chatbot, right, in a very simplistic way, getting reminders in terms of their medication time and all of that. It kind of eases things for the patients and you definitely would see a higher retention of patients. Again, I would relate this back to time to patent, right? In terms of how quickly can we get to patents, right? So this way the revenue starts. So there are multiple, and the other use case that I can talk about is how can we actually generate documentation for regulatory submission, right, leveraging gen AI. So those are some of the use cases that come to mind. Great. Great. So I mean, any of those, right, if we can implement those very well, it is going to significantly help with  bringing the time or shrinking the time for the regulatory submission and drug approval. Yeah. Yeah. What about on the manufacturing and supply chain side? What are some of the big impacts? Yeah, manufacturing and supply chain, definitely there are use cases, from a golden batch perspective and from even some of the traditional industry 4.0 models, right? Can you talk a bit more about that? So, from a standard industry 4.0 model, having digital twins, for example, one example where you have a digital twin for a manufacturing plant, that could be one use case. And now when you look at personalized medicine, and especially you talk about CAR-T technologies, Chimeric antigen receptors, T cells. Where in a person's blood is drawn. Here the manufacturing is very innate, right? I mean, in terms of drawing a cancer patient's blood and actually modifying the T cells and putting it back now requires a lot of process that manufacturing processes are tailor made for that particular patient at that point in time. Here's where technology plays a role in terms of providing the right information at the right time in that entire cycle of that process. Those are some that actually come to mind.

Can we go a bit deeper into Digital Twin? Because we have never talked about Digital Twin on this podcast. So if you can explain what that is. First I wanted someone to explain that to me. That was pretty interesting. So it would be great if you can explain that for the audience. Digital Twin essentially is creating a digital equivalent of a system. Now, it could be a manufacturing system.

You create a digital blueprint of it that actually lives by and is synchronized with the actual system. That is something that you can actually use from a prediction standpoint in terms of how in the future it is going to behave. Now, the exciting thing about this is people are also talking about digital patient twins.

So with all the variables and all the technology that is available, how can we bring all that together and have a digital patient twin? Now this is actually the next level. There is exciting times ahead when you have a digital patient twin so that you can do even research. You can possibly use that for the first part of a clinical trial, leveraging digital patient twins.

Sounds a little bit like science fiction at this point in time. But technologies are available and that's what makes this exciting that over the next few years, we would start seeing actual applications of this from a drug discovery, clinical trial standpoint. Yes. It's just fascinating. I find it fascinating. So very interested in that area.

And now let's also going back to our use cases. What about what are you seeing on the commercial side at this point or even medical affairs like whichever you want to cover? Yeah, from a commercial side, I think it is predominantly around, you know, content creation, right? Okay. It is definitely around content creation, wherein content that need to be generated for a global company, the same content may have to be translated in various different languages and each one of them has to go through a country level medical review. Imagine having a system leveraging AI and gen AI, taking care of that at least to a fair degree. Of course, there could be human interventions required for human feedback. Over a period of time, the system perfects itself with no hallucinations. But that's an area where there is tremendous amount of return on investment.

And there could be other areas in terms of how do we use autonomous agents, for example, from a call center perspective, right? That's the other use case. Also, companies are also talking about can we have some kind of a virtual healthcare representative, where in, of course, in-person meetings and all of that definitely adds value. But in the meanwhile, if there is a digital assistant available to physicians to actually interact with them, get the information they need on a particular drug, that would be useful as well. And if that is very context sensitive and also uses concept of multiple autonomous agents coming together, that will be huge, right, for the pharma companies. So those are some of the use cases. Yeah, and content creation is both, right? Like creating new content, but also reviewing the content, as you said about with MLR, either medical, legal, regulatory review as well, right? So with creation and review of the content, all of that can now be hopefully done in a much faster and less expensive way, right? Yeah, yeah.

Now, there is a lot of interest in capturing and integrating real-world evidence, and you referred to that earlier. So how is that changing drug discovery development process, and what are you seeing in the industry, and how does it pose its own data challenges? Right. From a real-world evidence standpoint, we know the data sets. There is claims data, there is EHR EMR data, there is patient-reported outcomes data.

There is patient portal data, caregiver portal data, even social media to some extent. So there are multiple rich data sets that actually need a lot of cleansing and integration that is required to actually leverage it. What that will actually do, we spoke about value-based care earlier, which is essentially how can we get the best outcomes for patients at an affordable cost.

And with accountable care organizations and outcomes-based payment models, real-world evidence is definitely getting a lot of prominence. One way to look at it is how can we make this more persona-based? Because this needs to hit scale. I think real-world evidence is actually going to help value-based care hit a certain scale, which will do a lot of good. However, in terms of bringing all of this together, think a persona-based thought process might definitely help. Payers are looking at real-world evidence from a standpoint of can they pay by outcomes? Can the outcomes very clear based on real-world evidence? Pharmaceutical companies are looking at it basically saying, in a world wherein if it moves towards a outcomes-based payment model, which it  already is, how do I ensure that, how do the pharma companies ensure that they have an understanding of the real world outcomes of a particular drug compared to other drugs that are available in terms of the standard of care, right? So that's one angle to it. I also think there's a patient angle to it, which is not getting often spoken about, right? And that's something I feel that can hit scale. So the value of real world evidence from a patient standpoint in terms of understanding what a patient's treatment and outcome is, right? And helping one understand in terms of a treatment that a patient may be undergoing, what's the real world outcome of that? Because that actually is a persona that can actually drive scale across real-world evidence. So you mean like patients with like certain - that have like certain demographic or  so like, you know, comorbidity, like how they are performing, is that what you're talking about? It is a patient's understanding of the treatment that they may be thinking of getting or they are already getting in terms of what's the real world evidence that that treatment is actually, you know, of value. You know, so that is, you know, because patients can also get more educated about the treatment process. Yes. Right.

And with real-world evidence data. And I think that kind of turns things around because that makes it even more valuable, the real-world evidence even more valuable when we have all the different personas actually leveraging it and using it. Yes, yes, absolutely. we talked about the data strategy, but I wanted to go a bit more into in terms of how much of a challenge do companies face with data standardization, interoperability, and how are they addressing it? I mean, it's great to talk about, yes, they should put everything together and have the government, but what are the challenges they face? How are they addressing? Have they been on top of all of this? Yeah, so that's a great question, Amar. I mean, we see all different kind of spectrums out there. There are companies born in clubs.

And that's a different spectrum in terms of having the strategy right there on cloud. But there is a lot of legacy out there and it is not easy to switch on right away. That is where it becomes very important to be very intentional about what business outcomes that one wants to drive in what area of the value chain. Somebody might say, for the next couple of years, I want to focus on getting the house in order when it comes to drug discovery, because pipeline is a challenge and we need to solve for the pipeline challenge. Right. I think then that becomes a very focused mission in terms of understanding the data sets that need to come together to enable drug discovery, leveraging computational biology and all the things that we spoke about. Right. Okay. So that is a very clear direction and it is usually top down and it is business driven. Right.

And then that area gets the focus around it. But it is always important to basically, while working on that strategy, have the larger organization in the picture. And that's what I see a lot of these companies doing, saying, hey, we focus on this particular area because we want to solve for this. But at the same time, whatever we actually produce need to be extendable to the other areas of the value chain.

So that then we can get more value out of whatever needs integration and can be integrated across the different parts of the value chain. So we're seeing a lot of that. Okay. And so, and into the challenges you talked about, there's different types of data sets, but also the legacy is an issue. What are some of the other challenges do you see with the companies trying to put together this strong data strategy governance? Yeah. Training is an important thing.

Right? I think getting and companies are focusing on that. Getting trained to the latest technology. Like I said, we have a perfect storm of innovation. So it is going beyond the data sets that one is normally used to and looking at other data sets that actually act to possibilities and look at it holistically in terms of how it can be integrated with the other data sets. So that requires an element of understanding and training.

Leveraging of technology, how can we do more and more in silico? That is the other angle to it. Those are areas where  companies are definitely spending more time and energy and seeing the results. Got you. It's still a spectrum as you see, like there are some companies who have mastered it. Change management is also an important part of it.

And the initiatives have to be driven from the top down. There is a cultural shift in terms of how data is used. In fact, the whole idea of modern data strategy is actually a cultural shift. It's not data for the sake of data. It is actually the business outcomes that is actually driving the data strategy. And ultimately, it needs to be co-owned with the business. Yes. So let me ask you there -  so it's a cultural shift mainly in like the data and the IT organizations, right? To have like this kind of a data strategy for like with business outcomes and so it's like, how is that going? Like how do  like people who are not used to that, like, you know, how's that happening? Yeah, again, you know, like I said, it has, it is top-down driven. It is intentional. And that it is important to actually say, hey, here's the business outcome that we're trying to achieve, say, in this period of time. And hence, what should be our data strategy? What should be our technology strategy? So this technology and data actually become part of the process. And that is when this whole concept of data as a product to actually enable business outcomes comes into play. And that's the mind shift that I'm talking about. That is the cultural shift that I'm talking about.

Technologists and business come together and basically become that one team to deliver to outcomes. Companies that are actually able to do that are the ones that are more successful. OK. And as you talked about this data as a product, is there a company? So there is the former companies where they think about the data product and then there are  the vendor, the data vendor companies. They talk about like these data products.

I kind of see a difference in that. Can you elaborate a bit more about when the pharma company says data is a product versus when a vendor company says that? Yeah. So I think when I say data as a product, right? Yeah. And this also goes on, I would also refer to the earlier conversation that we had around data monetization. When people think about data monetization, the thing that comes to mind is exactly what you said just now, saying you got data suppliers.

So for them, they actually collect data, they harness the data and basically serve a product and then they monetize it right away. That is one way of monetization. The other way of monetization is an organization leveraging their internal data sets and even external data sets, bringing it all together and actually driving a business outcome that brings in revenue. Now, that is the other angle of data monetization.

I'm talking about this angle of data monetization, where then data becomes a product, wherein you're actually serving that data as a product to a business team, or creating it, co-creating it with the business team, and then the business team is actually leveraging it, or reducing time to patent by reducing the time to discovery by say X number of months, reducing the time to clinical trials by a period of time.

And that translates into revenue. And you always need to have a mechanism to capture that in terms of what that saved, because then that becomes a very good feedback mechanism in terms of the level of monetization that this entire combination of data and technology had, so that it's no longer a cost center. All of this then becomes a profit center. And that's the shift as well from a culture standpoint. And it's very interesting, as you're talking about, is that data monetization, the definition of that in the pharma industry versus what's with the data vendors is completely different, even though we just talk about the same terminology. That's really interesting. Now, as you talked about, so now business top down, all of that is involved. And you did refer to like, the return on investment, like reducing the time to patent. How is that actually happening now? Because what are some of these strategies that the data teams or the IT teams are using to say, is like the dollar value that we're gonna put to what we're doing and for the data sites? What's happening right now around that? Yeah, so companies are definitely thinking about it.

Because when such huge investments are being made in terms of both data sets, harnessing those data sets and the technology layers on top of it to deliver to an outcome, there is always measurements. It's not always easy to measure it, but there are ways to measure it in terms of how that is actually based on even previous research, right? You always have a baseline to compare against and basically say, now that we have all of this, are we able to reduce the cycle time? A classic example would be, are we retaining more patients in trial compared to some of the previous trials? Because we put technology out there. Are we creating the clinical trial protocols in a fraction of a time compared to what we were creating before, hence start the clinical trial process quickly and put a dollar value to it, assuming that you put a successful drug out in the market after the clinical trials. So these are comparing against certain baseline, not always easy to draw a baseline, but there are always ways to arrive at an approximation of a baseline and compare against that and see whether one is doing better than before, leveraging all of this. So I like the idea where it's basically for those specific problems, you're defining, okay, well, this is how the metric was and how the metric is gonna change rather than going too big, right? Saying, okay, well, for this specific problem, this is what the value is gonna be. Okay? Okay, that makes sense, yeah. And you talked about this perfect storm. Where do you think all of this is going over the next... I'm even afraid to ask about five to 10 years. That just seems like too much of a window. Let me ask, say, maybe over two to three years, where do you think all of this is going?

We're definitely going to see more data coming out from all different kinds of data sets. And that's a good thing, right? Because we do have the technology to basically harness it. We're going to see increased use of AI and gen AI on top of the data. And we're going to see actual business outcomes even more than what we see today. And our use cases are going to become complex. And that's a good thing.

Because we are actually letting technology handle as much complexity as possible. And that's what I'm expecting will happen in the next two to three years. I'm extremely optimistic about what's out there. We have significant tailwinds right now, given all the different drug platforms that we discussed, enabled by technology and the technology at hand and where the technology is actually going from here. I mean, today, we are talking about digital patient twins, which is beginning to happen, but probably when that hits scale, you're seeing clinical trials that are happening first on digital patients and then actually going to actual patients, right? And that also shortens the time to patent, right? So there are a lot of tailwinds out there especially when it comes to pharma company and very excited about it. Great. Kannan Raman, Head of Healthcare and Life Sciences Delivery Professional Services for North America for AWS. Kannan, thank you for your time today. Thank you so much. Thanks, Amar. It was great talking to you. Well, Amar, what did you think?

Yeah, it was a fascinating discussion about data strategy, about a lot of the different use cases that we're seeing from data and Gen.AI right across the value chain. So a lot of the movement that's happening in the life sciences industry, it was good to get an overview of all of that. You know, one of the things he talked about was this shift in data strategy and how it's becoming based on business outcomes. So someone who is not intimately involved in data strategy. In some ways, it was surprising just to think that it wasn't always based on business outcomes. Can you talk about that shift? Yeah. So in my professional life, I've seen a few years ago when I first actually developed data strategy, it was all based on business questions and business outcomes. And that was actually very different from what the companies have been used to. It was definitely an innovation that  I thought would be, it should be obvious, but a lot of so far, the data strategy from the companies have been driven from the data or the IT organizations, which may not be very closely connected to business. But this is a very welcome shift that I find, you know, spending most of my time as a scientist or in commercial and then working on data analytics on that. I find that this is a very welcome addition. This is how it should be done, the data strategy, because when you are in a pharma company, all the data you're doing or all the analytics that you're using, that all needs to be to answer a specific business question, whether that's on science side or the business side. So it's a very welcome addition. Can I also talk about data as a product and monetizing data?

How are data strategies involving to do that? Yeah, so with data strategy, when you define different data products, as Kannan explained, it's about what is the outcome you're getting, how you're moving the needle from where the business outcome was before you use the data versus after that, measuring that. I still find this is still new for the industry, but I'm just glad that that we are getting there at this point. I think you also asked about this cultural shift. So it's a cultural shift, especially for the technology organizations in the companies. And I've seen that firsthand. I've been coming from more of the business organizations to look at how that is changing the way people have to think. But it has been a significant shift in terms of how people are now approaching things and how the tech and business are coming more and more together within a pharma company. A lot of times we think about the big AI applications that are going to be super transformative. You talked about agents which are doing these kind of routine daily tasks in a life science setting. You also talked about clinical trials and one of the cases kind of talked about was an example of taking the burden off of patients to improve retentions and studies. On the commercial side, he talked about content creation rather than, say, strategy or customization. People want to consider these big and bold AI applications, but there's this ubiquitous transformation going on. Do you think people are missing the depth and breadth of AI in the life sciences today? Well, see, I think a lot of people talk about many of these use cases. But what I've seen is that when people present, especially Gen.AI as this shiny new object, then they're talking about, OK, well, it is going to change drug discovery or commercialization and making these bold statements. The question is how? I it's not like just with one application that it's going to make that change. A lot of people, I think, especially people who are practitioners, who are working on the ground, they are seeing that there is not just one Gen.AI application that is going to change the world. It is going to be a lot of these specific business problems that you have. And there are so many use cases. I mean, in every area, there are hundreds of use cases that you can think about. And then Gen.AI can specifically solve those use cases. But then the question actually becomes, OK, which of these use cases, of these over 100 use cases, which ones do you want to focus on? Which ones do you want to prioritize on?

And that's something where I see companies struggling right now. That's why I kept asking the question of, what are some of the biggest use cases, most impactful use cases you see in the value chain? Because that's where the company should focus on. And he also said, it all needs to be intentional, right? So intentional in the sense, OK, focus on these, the big, high impact use cases. Let's focus on those.

But the way I see it is that a lot of times when you read about the Gen.AI, you're reading at a very high level, like, okay, well, high level use cases. But what's gonna make a difference is you come down a couple of levels and this is where the change is going to happen. This is where Gen.AI will actually work. This is where you don't necessarily say that it's a magic bullet that's gonna save everything. You're working on specific use cases, business problems to solve them. So it becomes very real, it becomes very measurable, rather than when you stay at the high level. The discussion went full circle and came back to the notion of outcomes driving data strategy and this being a cultural shift. Do you see evidence of a cultural shift taking place? Absolutely. I think there is definitely a cultural shift. What I've seen now is both in data and like the tech organizations where there is this - it's becoming imperative for them to work with the business teams much more closely to drive the tech strategy. But I've also seen a lot of cultural shift now with the business. By business, I mean the commercial organization or even the scientists or the R &D organizations, where there is a lot more awareness that we need to use a lot of data. We need to use AI to really accelerate whatever we are doing. And that is an awareness. I don't think there was that if I go back even 10 years ago. That awareness is here. And now what I'm seeing is, especially with Gen.AI, people are very much aware because everyone has been using a lot of these Gen.AI tools. And they are very much keen to see how that can be used in their business aspects as well. So a lot more awareness right now. think we, what I'm a bit afraid of people sometimes just like trying to go for use cases that are not going to be very successful. So I don't think people necessarily have the sense of, well, what is it that they can do easily? What is it that they cannot? And that's when we also talked about the Gen.AI use cases framework that I created. That was to address the question, okay, well, What are the use cases where the Gen.AI is much more feasible and is adding a lot of value versus not? I think that's where people do need to spend a lot more time on instead of just relying on some vendor to just come and do everything for them. So I think there needs to be much more awareness of, or some expertise around these making decisions about what should be the right prioritization.

It was a great discussion and certainly a lot to think about here. Amar, looking forward to our next one. Until next time. Thank you, Danny. Thanks again to our sponsor, Agilisium Labs. Life Sciences DNA is a bi-monthly podcast produced by the Levine Media Group with production support from Fullview Media. Be sure to follow us on your preferred podcast platform.

Music for this podcast is provided courtesy of the Jonah Levine Collective. We'd love to hear from you. us a note at danny at levinemediagroup.com. Life Sciences DNA and Dr. Amar Drawid, I'm Daniel Levine. Thanks for joining us.

Our Host

Dr. Amar Drawid, an industry veteran who has worked in data science leadership with top biopharmaceutical companies. He explores the evolving use of AI and data science with innovators working to reshape all aspects of the biopharmaceutical industry from the way new therapeutics are discovered to how they are marketed.

Our Speaker

Kannan Raman is the Global Head of Sales – Professional Services for Healthcare & Life Sciences (Global Accounts) at AWS (Amazon Web Services). With extensive experience in enterprise IT transformation, he leads strategic partnerships with the world’s largest healthcare and life sciences organizations. Kannan is known for driving cloud adoption, modernizing digital infrastructure, and enabling data and AI innovation through tailored professional services engagements.