IDEA Collider | Jackie Hunter

IDEA Pharma: IDEA Collider Jackie Hunter 

https://www.youtube.com/watch?v=K59qefVIsbA

Interviewer: So welcome to another series of IDEA Collider with leaders in innovation. I'm delighted to be joined by Jackie Hunter of Benevolent AI. First of all, would you like to tell us more about you and about Benevolent?

Jackie Hunter: Yeah, I'm a board director of Benevolent AI. I've been with the company since 2016, where I came to set up the drop discovery and development arm of the business. It was founded in 2013 and the vision of the founder, Ken Mulvaney, was really to harness the power of all the information, that it's unfair to make better decisions, come up with improved targets, and reduce the cost and time of drug discovery and development. [inaudible] big and small pharma. You know, there are big issues, big pharma is slow to make decisions, because of just big organizations usually are, but has a lot of money. And the smaller companies could be nimble and quick but didn't have the resources and the cash of the larger companies. And we thought there was a real opportunity to be able to improve the way we did drug discovery. I mean, it costs 2.5 billion to bring the new product into it to the market, there's no way that all the products that come to the market will report that. And the reason it costs 2.5 billion, is because you're paying for the failure. And I think it's probably the only industry I know that has a 97% failure rate.

Interviewer: Yeah. Interesting. I literally wrote a paper about that last week on the cost of failure and what you can do about it, so I'm going to dive into that, if that's ok.

Jackie Hunter: Yeah.

Interviewer: So even unpacking what you just said there, so there's an interesting thing that you said about better decisions. Decision making is typically poorly understood as a science in this industry. Can you tell me more about your thoughts?

Jackie Hunter: Well my thoughts are - I remember [Ken Keaton] Titan from [Thoughs] came to talk to, I was at JSK, and said, the one thing the pharmaceutical industry can do is make better decisions. Now that sort of sounds trite and a bit obvious. But I think the issue is that there's so much information around, that unless you use this new technology, you can't harness that information. And so, you're making your decisions on a very small portion of the evidence. And so, I've sat in senior meetings, in pharmaceutical companies, the person who was pushing that particular idea, that particular product comes along, they have to be passionate about it, because you need to have a champion. But you also need to have the evidence before you, to make a judgment to say actually, if what they are telling me that some of the whole picture. And you can use this technology to go for example, dive into past clinical trials, pull out more information about why a drug failed, or whether there are potential particular subset of patients that have responded in ways that, you know, just haven't been able to do. So, I think that for me, it's making better decisions, because you got more evidence, you have believed that if you got a better evidence base, the decisions you make, are going to be more likely to be correct.

Interviewer: And is there a sense of anonymizing or dehumanizing that process as well, because people tend to come with their biases and priors?

Jackie Hunter: That’s a really good example. Absolutely. We came up with a set of hypothesis in glioblastoma, a very horrible, nasty disease, unfortunate, I have lost two friends to it. And one of those particular targets was a class of drugs that I had worked on, for pain and other indications in the past. And they haven't worked. Now my bias, even though I know zip the bag clear of glioblastoma, or quality in generally would have been too damn prioritize that target, just because I have a body experience with it in the past. And yet the machine had surfaced it as being potentially very interesting. And long behold the person who was working on that project, took the set of hypothesis to a collaborator, who independently had screened the drugs and guess what? One of the gnomes that worked targeted that class that I just have downplayed in my mind and brought my own bias to it. So, I do think this lack of bias, certainly, when surfacing and evaluating the bulk of the information is very powerful. Of course, bias comes from insight and experience. And so, then you can bring the insight and the experience on that much more unbiased set of hypotheses.

Interviewer: And you mentioned that you think large organizations are slow, maybe old fashioned in some of the ways they think, and some smaller companies are too small to have that. Would you see what your vision is part of, is closing that gap between the large and capable, but the poor making decisions and the small potentially...

Jackie Hunter: Yes, definitely. I think one of our advantages is that we're the only really company doing what we are doing, going all the way from really early hypothesis generation through to sort of phase two clinical trial. So, we've got that capability. And it's important because the people we have on the clinical side, can think about inputting into the types of target the work or the diseases we work on and the data we get can feedback in. And we can bring the translational medicine further into the pipeline, but also some of the earlier thinking to stay forward into selecting and structuring patients. So, you need to have kind of circular rather than a linear process.

Interviewer: It feels like a process that people need to move towards that instead of away from because it's hard to argue with what you just said.

Jackie Hunter: Yeah, I agree. That's one thing.

Interviewer: Excellent. So, I wanted to get back to this definitions of innovation, because clearly you will be identified as in the inventive kind of world, and your own background in innovation gives you a lot of insight. How do you define innovation?

Jackie Hunter: I think for me innovation is not just having an idea about doing something differently or making something new. It's about taking that idea. And then actually delivering on it actually applying it. It could be a social innovation, it could be something like an impact on micro finance, that have a huge impact societally. Or it could be a product service. Or it could be a business model innovation. One of the challenges, I think we found is initially when we had discussions with larger companies, is they were very used to just operating a service provider model, especially with tech companies. Actually, what we bought is because of the investment that we've put into developing our technology, we wanted to have a collaborative model where we could bring our disease, if you like, knowledge graph and our predictive chemistry, and the company could bring their expertise in that disease area and data. And you know, that's a really good collaboration, but it took some companies a bit of time to get their head around it.

Interviewer: Why is that do you think? Is there a certain problem they have in decision making?

Jackie Hunter: Well, I think it's just it's doing things a bit differently. And I think if you think about the farmer business models, they tend to be quite stereotyped. And it's actually really interesting. There's been a lot of talk over the last decade about open innovation in the pharmaceutical industry, but I would say it's only in the last five years that companies have really become to adopt some of those open innovation principles. There's been a lot of in-licensing into a company, but a very little spinning stuff out, and that's starting to change. So, innovation is not only about having control of your idea to make it happen. It's about saying, this is the best way that we could make our idea create value for us or society or patients. And it may be that it actually goes through partnership with somebody else. I mean, we have a huge knowledge graph with billions of facts. And we can use it to generate hypotheses for any disease, but we can only work on a few diseases ourselves. So, to optimize the value for the company, you know, we don't just sit on diseases that we are not going to work on, we try to actively partner. So, we saw recently, this week, we announced with Miracle that we're going to collaborate on [inaudible 00:09:54] on chronic kidney disease and idiopathic  pulmonary fibrosis. So, you know, that's a language we can put our technology out there, if you will work collaboratively, to be able to get value for patients, get value for big collaborative companies and for ourselves.

Interviewer: So, there's some companies who are thinking about doing things differently, whilst doing the same thing, harder and bigger? 

Jackie Hunter: Yeah. 

Interviewer: And your distinction was between invention, you know, the kind of ideas and that kind of delivery of value, about the organization?

Jackie Hunter: Yes. I mean, anybody can have an idea. But if you don't move it, that's not innovation. And actually, I've seen some small startups where the founders are so worried about losing control that they don't take in the investment that they need. And so, the whole thing just drags on and on and on. Whereas if they've just thought, actually, we're going to do this, to do it quickly, we need this amount of money, we're going to have to give up some control, the whole project would have moved ahead much more rapidly.

Interviewer: So, do you see yourself here as being in that learning process as well? Because you're not fixed on your model either?

Jackie Hunter: Yeah, we're learning all the time, because what we're doing is so new. And one of the reasons that it's very good [inaudible 00:11:15], because her background is scaling up in tech brings to the Benevolent group of companies, a great wealth of experience. So, it's a learning environment for us all. And the other thing that we would really have to learn about is how to work together. So internally, a data scientists views the world very differently from a biologists or a chemists. And, you know, the biologists and chemists have to focus down on what questions they really want answering. And the data scientists have to be able to deliver all that, working with actually quite messy data in many cases. Biology data is messy.

Interviewer: Yes, apparently. And that was part of the belief of five years ago, was the farmer was missing a whole bunch of data scientists, and even interested in learning about how to play with those folks. So, you are kind of pushing that edge.

Jackie Hunter: Yeah. And I would say it took a couple of years before that really embedded. And, you know, now we have cross functional teams, we take the best of both worlds, really, we do have startups and stand ups and you know, sprints and cycles, but overlaying with that to some of the project management expertise that you see in pharmaceutical companies, so that we can, you know, both look to the longer term, because programs are long. But also maximize the kind of waste working that agile startups.

Interviewer: Yeah. Just to loop back to something you said earlier about your kind of cost of R&D and [March]. Is that part of your active vision to lower the cost of doing this?

Interviewer: Yes. And we've already shown we can because we can get to - and it's not us actually, other companies have shown that you can reduce the time from a chemical starting point to a candidate down to somewhere between 12 to 14 months. Traditionally, it's usually about three years, sometimes it's two, that straight away, you know, you're probably making [temp set] molecules, you are cutting the time by third, two thirds. And so, what you're actually going to see is, even in the early phase, you can cut costs quite considerably. And then if you look at some of the wok in patient stratification, if you can pick the right patient, then you should be able to do smaller and more effective clinical trials. And we've already mold some of that on publicly available data, for example. So, I'm very confident that we will increase the success. But I know we can cut the cost of failure.

Interviewer: And the paper I mentioned last week was on the failures and phase three, which is where we'd expect not to be losing drugs for advocacy, for example, you expect not to be losing them for, you know, those kind of variable things, like finance and strategic businesses. People are taking drugs in the face cream and stopping once they've bought the problem there is that they get the money back when they do launch another drug, and it becomes an expensive one. So, this kind of upward spiral as part of

Jackie Hunter: Yeah. I mean, the statistics show that you're still seeing 50% of jobs in phase three, failing, and that's just not good. 

Interviewer: Absolutely. This is what you would have to regard as predictable parameters.

Jackie Hunter: Yeah, so one of the things I think, by better understand the patient population, and this is I think why [inaudible 00:14:52], they have such a wealth of patient data, they could look to see how you could decide a face to try to mimic or even phase three to mimic your real world prescribing practice, as opposed to you've got very narrow criteria in phase two, you expanded a bit in phase three, and then it goes out into the wild west, when you put it out into the hands of your general practitioner. 

Interviewer: Yeah. And that is this kind of unspoken problem for the industry, which is we looked at was pure signal in the clinic, and then sent it to a very dirty world, where people are taking all kinds of things, or they are not taking them.

Jackie Hunter: Exactly. I mean, you know, people are on combinations of drugs. And we give exactly the same dose to a 50 kilos lady, as we do to a 120-kilo bodybuilder, you know, like sumo wrestlers. So, there's a lot of variation in general practice awaiting in specialty is that it just not accounted for at the moment. But hopefully, if we can look at the data in a much higher automated way, we'll be able to pick up the patterns of the signals that might allow us to cut that phase three failure rate.

Interviewer: How does it begin that feedback... because a lot of people don't want all that data once it goes into the real world, to come back. So, innovation itself, is that the kind of active process here? Do you actively manage innovation?

Jackie Hunter: Well, I think any company should be actively managing innovation, but it actually matters much more in a company like ours where you know, you need to get to the next value inflection point. So, we've got to make sure that we are innovating. And, frankly, if you're not measuring it, then that is bad for your business, because Einstein said the surest way of going backwards is standing still. So, we've got to be constantly innovating. Now, if you're looking at something like pharmaceutical industry, we can measure that in the number of patents filed. But in the tech world, patenting is less common, because it's really you will know how, and then we can measure our innovation, by the number of quality hypotheses we generate, as a surrogate for progression into things. And also validate. So importantly, the ones that we validate in our disease models, either internally or externally. And one of the things that I think we've been very good at is also designing the system so that when we make changes, we can check whether or not we have actually improved the system. Now I don't know, I think that's something that is routinely done. But certainly, one of the things and one of the benefits, I think about having clinical studies within our company is right from the get-go. We're thinking about version control, regulatory submissions, so that we have to make sure that everything is really very well documented.

Interviewer: So, decision quality is part of your metric, which is interesting, because you know, you look at kind of archaic approach with TPPs and everything else in the kind of total gates the [pharmaceutical] industry .There's actually very little reexamination of whether any of that is useful. And is that, like what you're saying here, is this embedded?

Jackie Hunter: Absolutely. It's embedded in our company. And I think, I mean, I remember when I left for SmithKline [Beach], and we went to look at some upside-down antidepressant, and postpartum depression. And I said, why? And then we should come back and say, why? They are being excluded, there's no documentation. Again, somebody's V point, well, and another classic one actually. When [Blackstone] and SmithKline merged, JP Guardian got migraine. And he didn't believe that you could tell that you could - he didn't get symptoms alerting him to the fact that he was going to get a migrate. And so, he didn't believe in migraine prophylaxis. But migraine prophylaxis has a huge market. Again, it shows have one person's view that actually shaped the decisions and the organization. 

Interviewer: And there's a lot of that, right, which is the people usually have one kind of views of disease to inform them. There are these strategies. Ok. So, you know, constant measurement, with internal references, external references as well?

Jackie Hunter: Well, we teach you benchmarking. Yeah. So, we have our internal benchmarks, and external benchmarks as well, saying the publicly available data sets, etc. So, I think all of that is really, really important. And we can also go back to the source of the information as well. So, when we generate hypotheses, we can look to see where that information surfaced, where the highest - you can go back through them, sort of 400,000 pieces of information if we wanted to, but it's really the key pieces of information. So, if, for example, it turns out that looking at that paper or that patented, it had flaws in it, we can signal down the system, so that it gets down, as a source of information later.

Interviewer: So, waiting is the hardest part. So, you can contrast large pharma, from your own experience with, you know, with the way things are today. So, what have you learned, over that sort of time period?

Jackie Hunter: Well, I mean, clearly, the importance of evidence-based decision making, it's really important. And actually, funnily enough, you know, I think that the whole thing about turning down to the cause analysis is really important. The trouble with us a scientist is, when we say problem, we jump into solution mode. And actually, quite often, the solution we're proposing doesn't tackle the root cause of the problem. And if we think about diseases, you know, finding what is the critical node in a particular disease pathway is really essential. But the other thing that's really important is people, you are only as good as the people in your team. And you need to have diverse teams. And I don't just mean in terms of protective characteristics, I mean, in terms of thinking. Quite often, the person who is the most innovative is not necessarily the person who's going to be actually inducing it to practice or managing the people who are going to run the project. And you've got to find a way to balance that and make room for those people, I should say this is one guys used to have, called it a group discovery leprechaun sitting on his shoulder, because he's always somehow just managed to pull out the best targets and come up with some really interesting ideas. But he wasn't necessarily the best manager of people. So, he couldn't be rising up in the organization, to be a leader, that is more senior level. But you had to find a way of making that person feel valued and wanted and giving them the space to do so.

Interviewer: So, do you think it's likely that you will, as you grow, you will end up looking like an organization, that we currently recognize what you think because of those things, that you look different?

Jackie Hunter: I don't hope we would be different. I would hope that, you know, we will change and evolve, and our business will change and evolve as we scale up. And I don't think you can be prescriptive about what that's going to look like, technology is changing everything so fast, it's how we work. You know, even now, we have cross-functional teams, we've been doing a chemistry design program with somebody in New York or Antwerp or Cambridge and London, virtually with a machine? Yes, it's just a different way of working.

Interviewer: Because it's a significant contrast between, say the tech careers grow, which is encouraging diversity, encouraging team-based collaboration, but not having you progress in your career to the point where you get to make all the decisions, that you've only been preparing for before.

Jackie Hunter: Yes. For me, it's been a bit of a learning as well, sort of seeing how cross functional teams that self-organize, can work together so well, it's not a structure that I'm used to in the pharmaceutical industry, where it's much more kind of regimented and rigid, but it's good feeling comfortable sometimes.

Interviewer: Well, that's one of the things that get kind of crossing of experienced, and the willingness to be wrong. So that doesn't occur very often. So, one of the things that matter, you mentioned culture, diversity, the ecosystem, decision making, what are the things that you're trying to model as you grow?

Jackie Hunter: Well, obviously, to grow, you need to have funding. And so, you know, we've been very successful in raising money, but longer term, we want to be able to show more proof points as a technology. And that's really why I came to work at Benevolent, because I knew we had to do drug discovery differently. And it's important for me, that actually we do discover that sense for patients, we're working on diseases, [inaudible 00:25:09] really underserved. I mean, you know, they are, unless you're someone like Stephen Hawking with death sentence, within [inaudible 00:25:21] are frequently a year. So, I think this technology is the only way that we will be able to solve those problems. And I'm really looking forward to the day when Benevolent can say we've got a drug in the clinic for ANS, we've got a drug in the clinic for [inaudible 00:25:41]. 

Interviewer: You are not far away. 

Jackie Hunter: What have we got? We've got a drug program in both those areas. So, I hope it's not that far away.

Interviewer: Not too far away. And what drives you personally? What's you motivation on Monday morning?

Jackie Hunter: My motivation is exactly that, is for this company to show we can do things differently and to produce effective medicines. And hopefully, through our collaborations with pharma partners, allow them to be more successful too. So, we can reach a greater range of diseases. And actually, I think it's this kind of thing that's going to save the industry really.

Interviewer: It's hard to see that it hasn't future in industry. What are the barriers to this becoming more part of things quickly?

Jackie Hunter: Well, I think there are - there are a couple of big barriers, I think the first is the availability of the correct talent, because data scientists are in high demand. And then having the environment that allows them to thrive and have the impact on the organization. And one of the things I love about [Henna], is it kind of puts the information in the hands of the scientists, whereas you've brought organizational setups in big companies where you have your chem informatics, your bio informatics, and they're almost gatekeepers to the data from to the scientist. And so, finding a way in which those expertise and skills can be utilized and still work, but at the same time, allow the move free flow and input the scientists onto those processes. That's what a lot of established organizational structures that I think could be an impediment to the adoption of this technology. And it's not just in companies, I think it's the same in universities as well, they organized a lot of very traditional lines, especially medical schools, I think we might have to think about

Interviewer: Blurring those lines?

Jackie Hunter: Yeah.

Interviewer: Some people are going to not do well in that, in the new setup. 

Jackie Hunter: Yeah, we desperately need the people who can straddle both the data science and the domain expertise, because they can facilitate the interaction and collaboration.

Interviewer: Ok. And I realize that we're using our time really quickly here. But there's a few things I always ask everyone, which is books that you would recommend, that you've loved, that have made a big difference to you

Jackie Hunter: The best book on innovation is Making Innovation Work by David Davila, Epstein and Shelton I think it is. I am impressed, super book. And they eventually sent me a copy of it while at GSK. And then I think Startup Nation is really interesting study of how Israel and its ecosystem, how it works and anything, and has been since successful innovation.

Interviewer: What was the key takeaway from that?

Jackie Hunter: Oh, I think the key takeaway from that actually - there are two. One was that because everybody went into the military, they were superbly well networked when they came out of the military, with not necessarily people from their own specialism. And secondly, that Israel has a lot of French money. And that allowed the flow of ideas to happen much more easily.

Interviewer: Very interesting. And you in your next five years? Where's Benevolent? Where are you?

Jackie Hunter: Well as I said, I hope to have a couple of drugs in the clinic for these serious diseases. And I'm hoping to move from making [side] to learning how to make capital, I think. 

Interviewer: Ok. I was waiting for the metaphor. Ok. Fantastic. And one is easy, [inaudible 00:30:14] at home? 

Jackie Hunter: I think so, there is a limit, it's only for personal consumption.

Interviewer: Ok. And in terms of Benevolent, this is obviously a scalable situation. Is it all going to be Benevolent or is there going to be multiple collaboration kind of?

Jackie Hunter: We already collaborate a lot with people, and I'm sure you will see announcements over the next 12 months about new units. As I said, there is a big collaboration. We are developing the time new relationships, because as I said earlier, we can use our technology in any disease, and finding creative ways of making it as available as possible. Working collaboratively with big companies, large and small. And another organizations will be really good, I think. 

Interviewer: Ok. And this is one thing that you sign of, as a sort of lesson of innovation within our industry, what would you say to everyone?

Jackie Hunter: Oh, that's a difficult one. The [wordless] in innovation, it has to be about the people, pick the right people and get the right mentality to not only have the ideas, but drive them forward. 

Interviewer: Perfect. Thank you so much, Jackie.

Jackie Hunter: My pleasure. It's been really great talking to you. 

Interviewer: Time went so fast. Thank you.