play_arrow

keyboard_arrow_right

skip_previous play_arrow skip_next
00:00 00:00
playlist_play chevron_left
volume_up
chevron_left
  • Home
  • keyboard_arrow_right Biomanufacturing
  • keyboard_arrow_right Featured
  • keyboard_arrow_rightPodcasts
  • keyboard_arrow_right The Impact of AI on Patent Research and Analysis in the Pharma [Ben Coverdale]
play_arrow

Biomanufacturing

The Impact of AI on Patent Research and Analysis in the Pharma [Ben Coverdale]

Ben Coverdale August 14, 2024


Background
share close

Yan Kugel is joined by Ben Coverdale, PhD, a seasoned global account director at Patsnap with a strong background in the life sciences and innovation and intelligence sectors. Ben brings a wealth of experience and insights into how AI and LLMs (Large Language Models) are transforming patent research and analysis in the pharmaceutical industry.

The Role of AI and LLMs in Patent Research

AI and LLMs are revolutionizing patent research and analysis in the pharmaceutical industry. These advanced technologies are enabling companies to navigate the complexities of the patent landscape, extract key data points from patents, and drive advancements in drug development and research. Here are some key points about the impact of AI and LLMs on patent research:

  • AI tools streamline the process of patent research and analysis, saving companies countless hours and resources.
  • LLMs can interpret and summarize data from patents, making it easier for researchers and innovation teams to leverage patent information for their advantage.
  • AI algorithms can predict and suggest ideas for further research by analyzing expired patents and identifying opportunities for innovation.

Leveraging AI and LLMs for Innovation

Companies can leverage AI and LLMs to drive innovation and develop new solutions in the pharmaceutical industry. These advanced technologies can help identify expired patents that have untapped potential for further research and development. By exploring patent databases and analyzing expired patents with the help of AI and LLMs, companies can uncover valuable insights and opportunities for innovation.

The Future of Patent Research and Analysis

The future of patent research and analysis is driven by AI and advanced technologies such as LLMs. With the use of AI solutions, companies can overcome challenges in patent research and analysis and unlock new opportunities for growth and development in the pharmaceutical industry. The speed of innovation, not only in the pharmaceutical industry but also in companies like Patsnap, is unprecedented, and the potential for leveraging AI and LLMs is limitless.

In conclusion, AI and LLMs are transforming the landscape of patent research and analysis in the pharmaceutical industry. These advanced technologies are enabling companies to make informed decisions, streamline the patent application process, and accelerate innovation. The use of AI and LLMs is paving the way for the future of patent research and analysis, driving advancements in drug development and research.

Episode Chapters:

  • Introduction to Ben Coverdale (0:00 – 2:00)
  • Overview of Patents in Life Sciences (2:01 – 5:00)
  • Challenges in Securing Patents (5:01 – 10:00)
  • The Role of AI in Patent Research (10:01 – 15:00)
  • Legal Implications and Partnerships (15:01 – 20:00)
  • Advancements in Patent Tools (20:01 – 25:00)
  • Conclusion and Contact Information (25:01 – 30:00)
  • Closing Remarks (30:01 – End)

Podcast transcript:

Please be advised that this is an AI generated transcript and may contain errors.

00:27 – 01:05
Yan Kugel⁠: Welcome to our podcast episode focusing on the pivotal role of patents in the realm of life sciences. Today we have the honor of hosting Ben Coverdale⁠, PhD, a seasoned global account director at the Petsnap⁠ with a strong background in the life sciences in innovation and intelligence sectors. Ben brings a wealth of experience and insights into how patents influence drug development, the challenges companies face in securing patents and the transformative impact of AI solutions in revolutionizing patent research and analysis. So, Ben, welcome. It’s a pleasure having you on this show. How are you doing?

01:06 – 01:10
Ben Coverdale⁠: Very well. Thank you for having me, Yan⁠. It’s great to be here.

01:10 – 01:50
Yan Kugel⁠: Thank you. So patents are really fascinating, Although many people might not realize it, how important they are for the industry. And we will dive into the discussion about it and how the AI tools today develop this seed and how they influence the research and manufacturing and so on. And before we dive into this discussion, can you give some overview of what you’re doing as an account director? What does it mean and how is it all connected to what we’re going to talk today?

01:51 – 02:32
Ben Coverdale⁠: Sure. So, yeah, a bit about my background before I start charging into what PatSnap does. So I’m a scientist by trade. I started off doing a PhD in biomedical materials at the University of Manchester and that was sort of my introduction into how you can start using patents as an additional data source to the scientific literature that we’re very used to in the research field in academia. We were starting to make three-dimensional scaffolds for stem cell engineering purposes and as well as looking at the scientific articles that were there, the patents that were really exciting because

02:32 – 03:11
Ben Coverdale⁠: a lot of the source of innovation that you find is published in patents. If anyone’s serious about taking their innovation forward, you’ll find it there. So that’s when I first got introduced to patents. And that is where I made the jump into into Patsnap and started working as an account director for this company. So yeah, I’ve been at Patsnap for around 4 years now. I’ve got the privilege of working with many of the largest and the most innovative pharmaceutical and biotech companies in the world. And we help them to improve their drug development processes. We help

03:11 – 03:40
Ben Coverdale⁠: them with their discovery of new assets. We also, I’ve got the benefit of working with our own product team as well. So we get a lot of feedback and requests from how we can exploit new technologies or if something revolutionary comes into the field, we need to develop the tools to allow researchers and IP professionals to search those. So yeah, I work with our product team as well to bring these tools to life and help innovation along the way.

03:40 – 03:59
Yan Kugel⁠: Thank you for this great overview. So now that we dive into the patents world, Can you provide some overview on how the patents influence the development and the innovation of new drugs and how it drives the industry as a whole?

04:00 – 04:40
Ben Coverdale⁠: Sure. Yeah. So what I mentioned during the academia that you do when you’re looking at scientific articles, the other side is patents and that is essentially your right to manufacture and sell whatever you’ve made, whatever your invention is. So within the pharmaceutical industry, it’s such an intensive place. The innovation is so strong now. There’s so many different facets of the industry which are being developed right now. And you have to be really defined in how you’re putting your technologies together. So it’s a very crowded space, but there’s also a lot of room for innovation in there

04:40 – 05:17
Ben Coverdale⁠: because there’s just so much scope to go at. So yeah, it’s really driving innovation in this sector. You see it through, you know, even the last few years, like through the pandemic and the advent of the mRNA vaccines you’ll see. Before the pandemic, you know, we didn’t see much of mRNA research, but now we’re seeing there’s around 3 or 4 times more patents and publications after the pandemic now. And it’s led us to develop new and more effective therapeutics for various, various diseases and ailments. And it’s the same across the entire landscape as well. Lots of

05:17 – 05:36
Ben Coverdale⁠: different therapies in there, including things like antibodies or CAR T therapeutics, things for targeting cancer specifically. So whilst it is a very intensive space and there’s a lot of competition, there’s a huge amount of innovation potential within this sector. So yeah, patents drive innovation.

05:37 – 06:24
Yan Kugel⁠: Right. So there are probably a lot of challenges there, right? So there are millions of patents on different areas, right, in the world. And when you do innovation, you need to make sure that even part of your innovation doesn’t take a book somewhere of some button that somebody has filed in your country or if there are some collaboration between the countries. So what are the biggest challenges that industries face when they need to understand whether what they’re developing already has been done or it’s new or are they allowed to do it? So how does it work

06:24 – 06:25
Yan Kugel⁠: and what are the issues there?

06:26 – 07:00
Ben Coverdale⁠: That’s a very very good question and 1 that is probably the key question in our industry, because you’ve got a new drug, how on earth do you establish whether someone else has used it in the past, or whether you’ve got what we call freedom to operate in terms of being able to commercialize what you’ve developed. So the wealth of data out there is huge and it is a massive challenge. There are a huge amount of litigation proceedings and all sorts of infringement issues because of this crowded space. But that’s why PatSnap comes in. So PatSnap is

07:00 – 07:36
Ben Coverdale⁠: a very, very comprehensive database. It has not just the largest patent database in the world, but we’re extracting things like sequence information, structure information. We’re getting all sorts of drug pipeline data in there as well. Because we’re using AI to connect all these different data points together, we can help to reveal those key insights to help you to plot better R&D strategies. You can understand who’s patenting in a particular space. You can use the tools to understand whether you do have freedom to operate or if you are infringing, if you do need to change something very

07:36 – 08:06
Ben Coverdale⁠: slightly. And it’s a much more effective way of cutting through all of that data, identifying those key technologies that you should be looking at pursuing rather than the older approach, which was, let’s just take a therapy and see how far we can go with it. Cause you know, I’m sure you can imagine the costs of developing this therapeutics, taking them into clinic, and then suddenly realizing that, you know, actually this has been done before. It’s a huge, huge cost. So, you know, we streamline a whole lot of that. That’s 1 of the main challenges that we

08:06 – 08:38
Ben Coverdale⁠: have. Another 1 is the whole R&D process is very uncertain now. So many projects fail. We get to a point where how are you selecting the drugs with the most potential before you see a return on this? So you want to try and make sure that you’re patenting at the right time. You don’t want to just go and patent everything because it’s a very costly process to build a patent portfolio. But you also don’t want to do it too late as well, because someone else might take your idea. So understanding how trends are evolving and how

08:38 – 09:02
Ben Coverdale⁠: different competitors are moving into the space, how different technologies are being developed, it’s all feeding into the same sort of the same pot where we’re collecting all this information together to make sure we’re making informed decisions. And again, that’s how PatSnap work with R&D and IP teams to help assess what’s going on here. So That’s how we overcome some of the challenges there.

09:03 – 09:27
Yan Kugel⁠: Right. So was it a manual process until the databases came in and they are in the helpful? So basically, people, when they started to research, they needed to do a very manual search and to understand whether their molecule was already patented. So how has it been? Was it done so far?

09:28 – 10:01
Ben Coverdale⁠: Yeah, it’s sort of the same old story, which we have iterated on for what 20 years since our inception. So it’s like our CEO originally started, he got very frustrated being able to trying to search for for patents and inventions in articles all across the world. He couldn’t find anything that was all linked together. So we thought, right, well, I need to put everything together in 1 place. And that’s where Pat’s not happened. That was the genesis of the whole tool, was trying to collect everything together, make it very easy for people to search for patterns.

10:02 – 10:39
Ben Coverdale⁠: But since then, we’ve been doing this for 20 years now, and the last 10 of which we’ve been developing in the life science industry and leading the field in AI development. We’ve been using the database that we have for patterns to link it with other data sources as well, pulling in the most comprehensive pattern sequence database in the world, got the most comprehensive structure database as well. And by linking all of these tools, developing algorithms to extract key findings, as I mentioned, innovation is packing these patterns full of great data sources. And we can extract all

10:39 – 11:09
Ben Coverdale⁠: of these to help R&D teams, but also IP teams to formulate that strategy. So It’s that initial question that Jeff had. I want to try and find this data. But now we’re using it with more complex sequence information. We’re looking at chemical modified RNA, oligonucleotide therapeutics. It’s the same question just applied to a more complex. And we’ve gone far beyond the initial question, but it’s just being applied in a different context.

11:09 – 12:00
Yan Kugel⁠: Yeah. So it takes time. Those are steps. Right. So it’s great. I think the best businesses start with the individual needs. Right. So a person understands, oh, you know, I’m lacking something and the best solutions come out from such a necessities. So that’s a really great story about the history of the company. And so you mentioned that right now you can identify not only specifically, it really goes into the molecules and everything. So it’s very fascinating. And just to understand also the differences between research and academia to establish commercialized drugs. So is there a difference in

12:00 – 12:22
Yan Kugel⁠: the need in such databases for the different types of companies, let’s say CROs or you know different start-ups, academia and big commercial companies that sell commercialized drugs. So where does it all come together?

12:23 – 12:58
Ben Coverdale⁠: Yeah, so the first point, you know, what we touched on before, developing your own technologies, that’s hard in itself. You’ve got to try and understand if someone has done that before. But then in terms of the commercial aspects, so a patent usually grants you access to or rights to that whatever you’ve patented for around 20 years, this enables you to sell that drug. And then after that expires, there are various things that you can add to this. You can get regulatory and exclusivity extensions on this. But understanding how you can try and keep hold of this

12:58 – 13:33
Ben Coverdale⁠: exclusivity around the drug, that can also be quite a challenging thing. So there are a lot of the big companies out there who are trying to continue their portfolios, make sure that the drugs stay with them as long as possible, because they obviously want to monetize it as much as possible. But equally, there are smaller companies there who are looking to identify what we call generics or biosimilars when these patents go off pattern, when they become available to manufacture in their own companies, bring them into their own portfolio. And there are tools that we can use

13:33 – 13:51
Ben Coverdale⁠: and perhaps that can accommodate for that as well. So we’ve got a lot of data around the expiry dates and things like the exclusivity data there as well. And we can help teams to basically form strategies around how you can develop from a commercial standpoint as well, not just a brand new research and innovation perspective.

13:53 – 14:24
Yan Kugel⁠: Right. So what are like the stages, right? Of the patent applications where so much data is needed and AI helps to bring everything in the right order. So what are the challenges there when you need to apply for the patent and you need to collect probably a lot of information, and you need more technological tools. So can you elaborate on that?

14:24 – 15:05
Ben Coverdale⁠: Yeah, sure. So I think traditionally, patents or patents analysis was seen as a trade where you needed a legal qualification. You needed to be trained up in the field to understand how patents were written. And that is largely the same now because you do need to write these patents and to understand to get them granted, you do need those qualifications. But it sort of separated the R&D teams, the innovation teams, the business development people in companies. They sort of had to rely on these IP professionals for the specific knowledge. Whereas now what AI is enabling us

15:05 – 15:39
Ben Coverdale⁠: to do is to really dig into how IP can be used to their advantage so that we can make it much more easy to understand using AI to interpret what the patent means. So PatSnap has its own LLM, which has basically been trained on our own database. So it’s trained on patent data, obviously. It’s trained on all of our life science information as well. And it’s getting to the point where we can summarize key data sets, we can understand the claims of a pattern in a very clear, concise format. We can actually see how the pattern

15:39 – 16:15
Ben Coverdale⁠: is gonna be developed and even look at spaces for where the innovation could be directed in the future. We can start to give suggestions on that. And in the future, we’re even looking at doing things like patent drafting and using the tool to support the whole development process from the very early discovery all the way through to the commercialization, our LLM is going to be there to support that process. So yeah, where it used to be quite a siloed area, which was just for IP professionals only, we’re now opening this data set up to allow R&D

16:15 – 16:26
Ben Coverdale⁠: professionals and innovation teams to really leverage the data that we find in patents, use it to their advantage and make it super clear as to what’s being claimed in them.

16:28 – 16:43
Yan Kugel⁠: Right, so you allow people who don’t have a lot of legal knowledge to understand better and read into what’s going on on the surface, right?

16:43 – 17:15
Ben Coverdale⁠: Yeah. So, I mean, a key thing that’s changing nowadays is a lot of these, like I say, a lot of these people have the expertise in focusing on IP. But if your whole company, if you’re a large company, if your whole company is relying on a particular department, Often you get these requests that often lead to large backlogs. And if the researchers could do those questions themselves, it would save a huge amount of time in terms of defining a strategy. Now, that’s a large company. And if you’re a small company as well, it’s equally important. If

17:15 – 17:41
Ben Coverdale⁠: you’ve got 10 employees and a new biotech and you’ve got 1 guy who’s supposed to be focusing on patents, but probably if you’ve got 10 employees in the company, you’re going to be doing so much more than just focusing on IP. You’re going to be looking at all sorts of other things as well as part of your day-to-day role. So if you have tools that enable you to very quickly assess the IP landscape, you can dig into the key factors that are found in those patents. It just makes your whole organization a much more efficient place

17:41 – 18:17
Ben Coverdale⁠: to be. And if you’re saving time, you can do other things that aren’t sapping like traditionally, looking through patterns and trying to find it. So if a pattern grants you a whole load more time that you can do so much more with, whether you’re a startup having those conversations to get noticed by other companies or going into lab and developing those new therapeutics. Or if you’re a large company, strategizing properly and making sure that you’re prioritizing the key assets in your pipeline, time is very important and that’s what PASS not provide.

18:17 – 18:46
Yan Kugel⁠: Yeah. Can you give examples of scenarios where companies did something wrong when applying for patents or they missed something and everything just went downhill because of some issues that they had while assessing patents and doing it incorrectly and manually and so on. So what are the dangers here?

18:48 – 19:32
Ben Coverdale⁠: I’ll save people’s blushes from actually putting the naming and saving people. But there’s so many incidents you’ll see of high-level litigation proceedings or legal battles between companies because something hasn’t been properly assessed. They haven’t done their due diligence on a specific therapeutic or something. So mentioning no names, there was a couple of years ago during the COVID pandemic, 2 of the key players in there, they are engaged in it or were engaged in a very intensive legal battle. And if that had been outlined properly, and they’d have the correct tools to be able to understand what

19:32 – 20:03
Ben Coverdale⁠: the infringements were, what they could do about it early on, then we wouldn’t be in that situation. So yeah, that in itself, the legal implications of this, it’s very good to work this out beforehand Because you could use it as an advantage. If you find that you’re infringing on someone and you recognize it early enough, you could even form things like partnerships or collaboration opportunities. You could license in their portfolio to yours to strengthen your portfolio. You can use it as it to your advantage. It doesn’t have to be just, oh, we’re going to go to

20:03 – 20:16
Ben Coverdale⁠: the league, to the lawyers now, and it’s going to be a messy situation. You could certainly leverage it and make it a more fruitful relationship for both parties. So yeah, another reason that we can, or another way that we can support that.

20:17 – 20:45
Yan Kugel⁠: Right. Yeah. So the due diligence is important. So I guess there are a lot of cases where companies were not aware even that they were infringing and after it can be maybe after 5, even 10 years that the company that files some patent or part of the patent, then they have claim for the revenues. So there are such occasions and dangers.

20:46 – 21:17
Ben Coverdale⁠: Absolutely, yeah. And it’s an incredibly costly process. And it goes on for a long, long time. And again, as I keep coming back to that the time spent on these things, which is you don’t want to be spending your time in a litigation battle. You want to be spending your time developing therapeutics, finding the next the next thing, developing the new and the new invention. So yeah, all of these things contribute to having a more successful, successful business which allows to, you know, innovation to thrive.

21:18 – 21:40
Yan Kugel⁠: Right. So how do you see AI influencing the patent landscape in a way that it would not only make life easier to research, but actually drive innovation? So does it make sense that it would do it and in what way?

21:41 – 22:18
Ben Coverdale⁠: So the patent side of things and doing the classic FTO search, as we call the freedom to operate search. That’s just 1 thing that we specialize in. The other thing I mentioned, there’s so many different data types that are hidden within patents. It’s got so much information in there from like the biological information in there, the structural data, there’s processes, there’s experimental information in there as well. And 1 of the things that PatSnap is very good at is creating ontologies of this data, collecting it all together, categorizing it, and making it readable for the user. So

22:18 – 22:56
Ben Coverdale⁠: 1 example of this is we’ve curated a data set of 100, 000 antibody antigen pairs, which can be used to train in silico models. This data is very good in itself, but we actually are able to extract further information from patents as well. So not just the pairings themselves, but we can get things like epitope data, affinity data as well. And we’re collecting all of this together to make it usable for these teams who are developing like in silico models, so AI using their own AIs to establish these new therapeutics. So that’s just 1 example. There’s

22:56 – 23:06
Ben Coverdale⁠: loads of different ways we can extract key data types and make it usable for development and innovation beyond just the classic freedom to operate and strategy.

23:11 – 23:48
Yan Kugel⁠: Right, so patents are usually valid for 20 years, right? In total, And can the AI predict or give you some hints which expired patents you can use to leverage them to create new solutions? Is it something manageable that even if it’s not happening today that it can, the machine learning can be trained to do something like this, to search different patterns and suggest ideas for further research?

23:49 – 24:19
Ben Coverdale⁠: There’s no reason why not. And actually, as part of the tool that we have right now, it’s called Hero. So it actually works. It’s part of our LLM. It’s our AI tool. But 1 of the things that you can do with it is, as I mentioned earlier, you can summarize patent data. We use that by going through what we call patent DNA to assess what the patent was filed for, what are the key findings in there, and then what the technical outcome of that was, what the problem it was designed to solve. And as part of

24:19 – 24:50
Ben Coverdale⁠: that, we can use it in the context of the whole innovation landscape to understand what’s been addressed in that pattern, but what also could be developed that hasn’t been touched on in the pattern, So things that could be improved on, that you could file a new patent in the future on. And that’s something that we’ve released towards the end of last year. And that’s something that we’re developing all the way through this year as well as part of our enhancing the capabilities of the LLM. I mentioned there’s a few things in there as well, like looking

24:50 – 25:24
Ben Coverdale⁠: towards actual patent drafting. And there’s things like being able to like summarize the entire data set. We’re actually bringing out our own LLM trained on all the life science data that you can ask specific questions. You can get it to extract and summarize key information from like trial data and drug pipeline information. And then even beyond that as well, we’re actually looking at being able to use our LLM on client data as well. So not just PatSnap itself, we’re not just saying, here’s the data that we’ve got. We’re saying to our clients, how do you want

25:24 – 25:36
Ben Coverdale⁠: to use the data? Use our interface to interpret and extract those key data points that are relevant to you. So yeah, lots of things that we can do here to accelerate that process.

25:38 – 26:21
Yan Kugel⁠: Right, so you mentioned that it would save countless of hours, right, for companies big, small, right? So do you think this will support more advancement because you save time on things that probably take weeks, months, maybe years to research, right? So if you can do it quickly, can put more effort in the research and then also companies that probably started working on something that was already patented, then they can understand, okay, they can just move to the right or to switch projects, right? So there are so many things that are pivoted around patents that if you

26:21 – 26:33
Yan Kugel⁠: do it correctly, you save time, you can make way for innovations. But is there any other things that you can think that we didn’t mention that can support the innovation.

26:34 – 27:07
Ben Coverdale⁠: I mean, there’s a few new tools which I’m really excited about at the moment. Part of our biosequence tools. So I mentioned I’m a scientist and the biology side really fascinates me. But yeah, I mentioned saving time with processes that existed before, so you could read through patterns and try and understand the key datasets within them. There are other things as well that are just not possible to do so at the moment. And PassMap have come along with a couple of new tools. 1 is our chemical modification search, which is for oligonucleotide RNA therapeutics. You can

27:07 – 27:46
Ben Coverdale⁠: actually search for chemically modified patterns within those RNA strands to understand who has been modifying those particular bases for whatever reason, stabilization or functional benefit. We’ve also got a degenerate searching tool as well, which allows you to search for degenerate sequences. I don’t know what a degenerate sequence is. It’s essentially a sequence which can be claimed as anything, any base, which is very helpful. But it means that you can apply a lot of protection on those sequences. And it’s very, very hard to go and, firstly, find them, but also try and understand what they’re actually claiming.

27:47 – 28:24
Ben Coverdale⁠: But perhaps now we’ve developed with AI sophisticated algorithms that can break down these difficult to find sequences. And it just makes it so much easier for you. These situations where you just wouldn’t be able to search these previously are a new avenue to explore. So we can’t even compare the time saved before because it’s an entirely new process which we’re enabling people to do. And that sort of thing is really exciting because it’s opening up a whole new era of innovation for us and companies working in the biologic space. So yeah, that’s really cool. I really

28:24 – 28:25
Ben Coverdale⁠: like that.

28:26 – 29:03
Yan Kugel⁠: Yeah. So are there such things as patent graveyards where you would say, okay, there are patents where we’re filed, but nobody, maybe companies went by bankrupt or they didn’t pursue some technological advancements. So do you think companies could go to such databases and just try to understand the patterns that were not followed up on and just pick it up and just continue the innovation from there? Do you see like companies really doing something like this? Does it make sense?

29:04 – 29:39
Ben Coverdale⁠: You definitely can do. Yeah, it’s quite a big use case to be able to do that, to try and pick up technologies that went to a certain stage and then just for whatever reason weren’t carried through. And there’s a few other uses as well. You mentioned graveyards, but there’s also like examples of larger companies making these patent thickets, they call them. So it’s like just patenting across 1 drug will have multiple patents attached to it just to confuse the landscape. They don’t know what the weather picking it up. But Pat Snap again, the ability to extract

29:39 – 30:07
Ben Coverdale⁠: the key insights, the key information, where the core patent is, and be able to take you straight to that. It means it’s very easy to go and find those, whatever has been claimed, the key sequence, key drug of interest, and we can make those links and connections very, very quickly. So yeah, whilst companies can do that, they can go and search through like old patents as well. We can help you to pinpoint the key patterns. When you come to 1 of those tickets, you could be like, right, I know exactly what these guys are talking about,

30:07 – 30:11
Ben Coverdale⁠: what they’re claiming to help me develop my own portfolio.

30:12 – 30:55
Yan Kugel⁠: Yeah, that’s very interesting because I think that there are a lot of innovation that was happening in the wrong time. So we’ve seen it before, let’s say, with Microsoft, who invented a touchscreen 10 years before Apple, but there were no apps, there were no internet, so you couldn’t use it. And then suddenly it just exploded, right? So there is so much potential to go and take a look at older patents that might have not developed because there was no technology to complement it, right, to combine it. And now the, this, you know, can be done and

30:55 – 31:37
Yan Kugel⁠: things can be developed further. So that’s a highly interesting. And besides what we were talking about that you mentioned that there should be AI that would help draft and help connect the dots. So do you see any other advancement? If you look in the future, which you think is still not yet there, but I think it could happen. We might be working on this, or we might be working on this, or at the moment it’s just imagination. But if we could do that, that would change the landscape.

31:38 – 32:14
Ben Coverdale⁠: Well, there’s so much. And it’s really difficult to try and wrap my head around how much we’ve developed in the last year. We, and even in the last like 6 months or so, we’ve released, like I say, that degenerate searching tool, we’ve got our chemical modification platform. We’ve got our own large language model, which can, you can ask questions and it can interpret and summarize data for you. So the speed of innovation, not only in the pharmaceutical industry, but in PatSnap as well is unprecedented. I don’t know where we’re gonna be in, you know, 6 months,

32:14 – 32:51
Ben Coverdale⁠: 12 months time, But we have a huge team of developers who are very well-versed in how to make these tools and how to get the most out of this data. And I can only see it improving. There’s more and more data coming out there every day. We’re coming up with more innovative ways of extracting this information and making it readable for the user. So yeah, there’s all those things like pattern summarization, drafting, analysis, and tagging key drugs, targets, indications, pipelines, whatever you want. And even moving towards, like I said, using client data, using their own information,

32:52 – 33:02
Ben Coverdale⁠: using our LLM. So there’s limitless potential, really, for how you can use the data. So I’m really excited to see where it goes.

33:03 – 33:14
Yan Kugel⁠: So do you see that this explosion of LLMs suddenly changed the industry and just opened completely new doors for you, which you couldn’t have imagined, like, I don’t know, 3 years ago?

33:15 – 33:52
Ben Coverdale⁠: Pretty much, yeah. I would caveat it by saying, if you just throw out an LLM there, it’s only as good as the data it’s trained on. You know, if you, if you look at something like chat GPT, for instance, it’s an, it’s an excellent model for asking it questions, understanding the entire internet, and getting a response from it. But with that comes problems. If you start asking it more complex questions, you start asking it more in-depth questions about specific innovation or specific therapeutic, It will draw sources from multiple channels. And that’s where you get things called

33:52 – 34:25
Ben Coverdale⁠: hallucinations, which can often be contradictory. They can mean that you have to say, oh, hang on a minute. I’ve got 2 competing opinions here. I need to go and double check them. And then that whole thing I said about time, you have to go back again and double check. You have to validate which 1 is correct, find the original sources. It can prolong the whole exercise. So with PatSnap, So we’ve trained that LLM specifically on our own data. So we know it’s the most comprehensive patent database on the market. We’ve brought in the most comprehensive sequence

34:25 – 35:00
Ben Coverdale⁠: and structured drug information database in the market. By combining all this key data together, we’re able to make very confident assessment of however you’re trying to extract information from the LLM. So yes, it’s important and I think the LLMs are very good, but you need to make sure that the data that they’re based on is accurate before you’re trusting them. So for the purposes of what PatSnap is doing, helping in the pharmaceutical industry, development, discovery areas, that is an excellent data set to go off. So, Yeah.

35:01 – 35:27
Yan Kugel⁠: Great. Great insights into that. So I agree with you completely on this. So the data should be there. And the LLMs are as good as the data that they work with, right? So it’s really fascinating topic, Ben. So do you have any key takeaways before we finalize the talk, which you think we didn’t touch, but I think you think it’s very important to share?

35:28 – 36:08
Ben Coverdale⁠: Sure, I think what we’re seeing nowadays, moving into 2024 and the advent of AI, all the LLMs that we’re mentioning, it’s very obvious across the field. There is a decline in the return on investment that we’re seeing in drugs across the world. We’re spending more and more each year, I think it was $250 billion or something in pharmaceutical R&D each year. And we’re seeing diminishing returns on these year over year. It’s a hard time to be developing. And you can see that from things like the large companies are making large layoffs. It’s not affecting just the

36:08 – 36:49
Ben Coverdale⁠: small businesses. It’s large companies. We’re seeing a real paradigm shift in how we’re developing new technologies and innovations. And this is why we need to start using AI to become more effective at uncovering these technologies. So companies like PatSnap moving forwards to understand these key data points, being able to make those efficiencies. You can identify key information from the patterns. You can make much more confident decisions about the strategies that you’re making. Instead of taking 10 candidates forwards and spending a lot of money, you can really focus on that 1 candidate that is going to make

36:49 – 37:23
Ben Coverdale⁠: it to the market. Or as we’re seeing more and more now, using in silico models to make thousands of candidates and selecting the key ones that you want to take through your development and into the clinic. This is where the development is now starting to happen. So we’re seeing companies being left behind right now. Those who aren’t using AI, those who aren’t using tools like PatSnap to embrace innovation, we’re seeing them fall behind and they’ll be on the decline. The ones that are starting to use AI to their advantage, they’re the ones that we’re seeing to

37:23 – 37:54
Ben Coverdale⁠: develop these new therapeutics, these new innovations. And it’s not just the pharmaceutical industry, it’s across the whole innovation landscape. But obviously I’m biased and I love the farmer industry far more. But yeah, that would be my key takeaway is those companies who are starting to use these innovative tools and to be more effective as other ones that we’re seeing, we’re seeing succeeding. So yeah, if you’re interested in looking at how Patsnap can help that, drop us a note. Come and see how it works.

37:55 – 38:20
Yan Kugel⁠: Great, Ben. Thank you very much for those final words. So we will have the contact information for Ben Patsnap in the episode description. So feel free to reach out to Ben. And Ben, thank you very much for this great talk. It’s a very fascinating field and I wish you the best of luck.

38:21 – 38:33
Ben Coverdale⁠: Yan⁠, thanks. It’s been a real pleasure. I’d just say check out our website. We’ve got a great AI page there. Feel free to contact me directly. I’d be very happy to show you about what we could do.

38:33 – 38:48
Yan Kugel⁠: That’s great. Would you like to share the website address of the URL of a pet step just to, if it has some ending, which is unusual. So, pet step.com.

38:48 – 39:00
Ben Coverdale⁠: Search passnap.com, yeah. Or touch base with me, it’s just bcovidale at pasnap.com. And yeah, we’ll show you around, and come and see what we can do.

39:00 – 39:06
Yan Kugel⁠: Cool, Sounds very fascinating. I would take up on this offer myself as well. Cool. So thank you, Ben. See you later.

39:06 – 39:07
Ben Coverdale⁠: Thank you, Yan⁠.

Tagged as: , , .

Rate it
Previous episode