Yan Kugel is joined by Will Moss founder and CEO of Seal, to talk about the critical need for automation in GXP processes. Will shares his expertise on balancing automation with human oversight, addressing data privacy concerns, and the future of quality management in the pharmaceutical sector.
Table of Contents
Will Moss’s journey into the pharmaceutical industry began with a simple realization: the GXP industry is ripe for automation. With a background in chemical engineering from Cambridge University, Will transitioned into data science and software engineering before founding Seal nearly six years ago. His goal? To help pharmaceutical professionals streamline their processes and reduce manual labor.
Key Challenges in the Industry
Will highlights several challenges faced by professionals in the pharmaceutical sector:
Automation can significantly enhance productivity and efficiency in pharmaceutical processes. Here’s why:
Seal’s Approach to Automation
Seal offers a unique toolkit that empowers users to digitize their workflows without needing extensive programming skills. This approach allows companies to create their own automated systems, adapting and improving them over time.
As Will points out, the journey to successful automation involves understanding the specific needs of organizations. Here are some key takeaways:
Cultural Considerations in Large Organizations
n larger companies, the challenges can be more cultural and people-oriented. Will notes that:
Building Momentum
To foster acceptance of automation, Will suggests:
Will discusses the common fears that companies have regarding automation, particularly concerning data privacy and control:
Easing Concerns with Transparency
To alleviate these fears, Will recommends:
The insights shared by Will Moss on the podcast underscore the immense potential of automation in the pharmaceutical industry. By embracing innovative solutions and a strategic approach, companies can enhance their operations, reduce costs, and ultimately improve patient outcomes.
As we look to the future, the collaboration between technology and human expertise will be crucial in shaping a more efficient and effective pharmaceutical landscape.
Episode Chapters:
Please be advised that this is an AI generated transcript and may contain errors.
00:27 – 01:18
Yan Kugel: Welcome to the Quality Talks Podcast. Today we are excited to have Will Moss, founder and CEO of Seal with us. Will is a chemical engineer who recognized the need for automation in GXP processes and now leading to innovative solutions in the pharma industry. And in this episode, we’ll discuss the balance between automation and human oversight, data privacy concerns and how GMP professionals can effectively integrate AI into their practices. We’ll also explore future innovations in quality management and cultural shifts needed within the pharmaceutical organizations. So we’ll welcome to the podcast. So You are a chemical
01:18 – 01:39
Yan Kugel: engineer and you studied the engineering in Cambridge University and then you haven’t been in the industry before and then you came to the idea and to the realization that the GXP industry requires automation. So how did it come to be?
01:39 – 02:18
Will Moss : Yeah, for sure. So you’re absolutely right. So my background studied chemical engineering and I guess I was always excited about this idea of taking things from the scale at which the scientists just figured out, you know, the product, whatever that is, and figuring out how to scale it up, how to get it out to people. And so I actually went after studying chemical engineering, went out and became a data scientist, worked as a scientist, worked at software engineers and great companies. But the interest was always there. So I started this company nearly 6 years ago. And
02:18 – 02:55
Will Moss : initially, what I was doing is I really wanted to help people who I thought were working on really important products and sort of life changing, world changing products in my view. So things like cellular agriculture products, so lab-grown meat, and cell therapies were 2 things I was particularly excited about. And so I went to those companies and I said, look, obviously I don’t know the reality of the industry. You know, at the time knew little, but I did have my own set of skills, which were building software and using data. And so I spent a couple
02:55 – 03:27
Will Moss : of years building bespoke tools for all sorts of industries. So from process development all the way through to manufacturing and all sorts of scales, small biotechs through to some of the biggest pharma companies in the world. And I guess I saw the same or similar enough set of problems again and again. And I kind of saw that the more quality was involved, the poorer the solutions were that people had to work with. And that’s where we came up with Seal. And we thought there could be a better way to do this.
03:29 – 03:56
Yan Kugel: Right, So the idea behind Seal was to help pharma professionals automate the processes and less manual labor, right? So do you find that a lot of companies are still doing too much manual activities, which is outdated and there’s such greater alternatives to really expedite and save time.
03:56 – 04:27
Will Moss : Yeah, for sure. So I mean, I think the reason I’m so excited about the industry is there’s so many smart, really, you know, amazing, brilliant people working on products that are like this, if this, if you can get this to people, like it’s going to change, you know, thousands, hundreds of thousands of people lives. And then you go into the, the facility or the lab or wherever it is and you look and you see this person’s got a PhD, they’ve trained for like years and years and they’re like incredibly smart and they know these different things
04:27 – 05:15
Will Moss : and then you see the reality of their work and it’s quite shocking really. I’ve kind of stood there, watched and kind of note down timings and things. I’ve spent more time, obviously, as you get towards manufacturing, things become more efficient. But process development, I reckon 70, 80% of that time is like people not using, you know, the skills that they’ve developed. Instead they’re kind of running processes or checking things or double checking things or figuring out how to share something with some other team and there’s just it’s like it’s really shocking and I think people in
05:15 – 05:52
Will Moss : the industry are aware of that Like if you go to the biggest, most famous operations in the world producing stuff that’s huge, huge scale, paper is still unbelievably prevalent and that itself inherently, you know, there can be times when that makes sense. I mean, I don’t believe that, but I think that’s a real sort of sign that this industry hasn’t taken the leap forward that they have in many other industries. But that’s for good reason. Like, obviously, it’s going to be a conservative industry. We need to make sure that things are done right. But I think
05:52 – 06:15
Will Moss : my counter argument is there’s risk in leaving things as they are. There’s so much inherent risk in not moving forwards, not getting all your data together in 1 place, not automated things and having people do things manually. And yeah, enormous opportunity to change the way we work and get all these brilliant people working and focusing on what their skill sets are.
06:16 – 06:47
Yan Kugel: Right. So I cannot agree more with it. And as you said, it’s a conservative industry and it’s change resilient. Yes. A lot of the time, right? And I think that there are a lot of people who are also saying that there can be an over reliance in automation, AI, and then the human touch is lost and it can be dangerous. What do you think about this? Do you hear it a lot and does it make sense?
06:48 – 07:23
Will Moss : I think you’re completely right, Yan. I think when you look at how most sort of what people call automation projects appear and are done, they’re these huge projects where some sort of big budget’s been put aside, and then they’re like, we’re gonna automate this thing over here. And it’s months of work, huge program, the people who know the detail are involved in it. Maybe they’re brought in some sort of third party, some consultancy, some sort of experts in automation. And the reality, you know, the people who know what needs to be automated, the people who are
07:23 – 07:56
Will Moss : feeling the process, who know what they care about and what they don’t care about, they’re not in control. And that’s really what I wanna change is that as I come back to the same kind of idea is that we really believe that, I know that these are brilliant, smart, incredible people who work on wonderful things like if you’re much better rather than life science companies in particular going out to third parties for automation projects and sort of teeing up a budget and doing some sort of ROI and calculation and trying to justify this thing. And then
07:56 – 08:21
Will Moss : it never coming out quite how you want, not being in control of it, not being able to adjust it when it doesn’t work exactly how you want losing the human touch as you said, saying, you know, this isn’t doing quite what we want, or we have a new project now where we want to change how this automation runs slightly. Or let’s call up, you know, the consultancy report again and get them to rebuild and adjust it, spend another sort of hundreds of thousands of dollars. You know, I was speaking to a company the other day and
08:21 – 08:52
Will Moss : this is exactly what’s happened. They’re buying Seale because they spent, you know, half a million, and this is a tiny, tiny lap, tens of people on automating this 1 sort of analytical workflow around testing their samples and they brought in this consultancy to automate that and streamline that because it was taking up money and time and had all these sorts of problems but it’s just not worked because they keep every time they need a change they have to go back to that company, call them up, say, we need you to change these things. And like, the
08:52 – 09:17
Will Moss : feedback loops are far too long. And something you learn in the world of software is that if you can compress feedback loops, so when I’m building software, if something, if the feedback loop is at an hour, so for example, I write some code and then I have to wait for an hour to the system to build to see the results, see if it worked. That is so frustrating and it drives me absolutely mental and I can’t move at the speed I want to, but if I can get that feedback loop right down to like seconds, So
09:17 – 09:47
Will Moss : I write some code, see the result, know it works. I can work 10 times, 100 times faster, and I’m a lot happier. So kind of bringing these 2 ideas together, what we do at Seal is instead of saying, look, we’re a third party, we’re gonna come in and do stuff for you, we’re saying, look, here’s a toolkit. Here’s a platform where you can take your existing procedures, take your existing workflows, and digitize them for yourself exactly how you want. So over time, you can continuously improve them. You can go to version 2 of this automation, version
09:47 – 10:15
Will Moss : 3, version 4. You can take it and stick it in some other process. You can build it out into particular functions. And I think kind of the main idea that for us as Seel is like, can we rather than saying that we know best, we’re saying, here you go. Here’s a platform. Here’s a set of tools. You’re the experts, you’re bloody, bloody, you’re clever, you know your process is a million times better than us. Let’s give it to you and you build what you need for yourself. And if we can help, great, but ideally, you can
10:15 – 10:23
Will Moss : run with it and build amazing things and adjust your automation for yourself rather than having to get something to do.
10:23 – 10:28
Yan Kugel: Right. So that sounds very interesting. So then Seal is like Zapier.
10:28 – 10:29
Will Moss : Exactly. So it
10:29 – 10:36
Yan Kugel: basically allows the users to build their own workflow without programming skills. So is
10:36 – 10:47
Will Moss : that the- Exactly, yeah. The core idea rather than being some sort of just the automation, you can also put procedures in there. So you can have some SAP that you can digitize and then you can automatically when this finishes do this. Exactly right,
10:48 – 11:10
Yan Kugel: yeah. Right. And how well does it then connect with other tools that the company might have, like the, let’s say, LIMS or QMS? So do they need to make sure that they have API access to those tools or is there other data extraction options that are available?
11:10 – 11:43
Will Moss : Yes, it would be nice if everyone had this sort of open accessible API, but I think everyone probably knows in this industry that And it makes sense that it’s kind of conservative you end up in many instances With an instrument or a bit of software or something. It does the job. You know how it works, you know It’s got its inputs and its outputs, and you know that it does it. So why remove it? But obviously we need to figure out how to interface that. So yeah, if it’s a LIMS, easy to, modern LIMS, easy to
11:43 – 12:18
Will Moss : interface with. Other modern software, your MRP, ERP, if they’re modern, very easy for us to just use the API’s to connect them, whether that’s sending data from our platform or putting data out to the platform, whatever you need to do. So we see ourselves kind of as glue tying everything together. But in many instances, like instruments are offline. Things are done offline or the software’s outdated and doesn’t have an API. But in pretty much every case we’ve seen, like there’s this kind of voice, yes, we don’t get a real-time view, but often it’s okay. What we’ve
12:18 – 12:47
Will Moss : built is a way to kind of like Dropbox or Google Drive. You can put this little widget on a computer or Raspberry Pi and watch for when the CSV comes off. And so we can go, well, yes, you bought your ERP in 1990 and it does, it works for you, but, and it’s in some server in your room, but as long as you can auto set it up so that it drops a CSV or a file of any type somewhere, we can easily pull that out and put it in the cloud. And that’s, that’s fine. Ideally,
12:48 – 12:58
Will Moss : so the younger startups tends not to be a problem, but as we work with more mature organizations, it’s really important that we have a way to work with everything.
12:59 – 13:33
Yan Kugel: Right. So you still have, you need to have some tech savvy people working with those tools, right? They still need to have some understanding of how to design processes and feedback loops from the computer and from the system and to, you know, how to resolve bugs and then make sure that everything works, right? So who are the usual people who work with your systems, for example, are those CSV experts or computer engineers in the companies?
13:34 – 14:07
Will Moss : Yeah, so we’ve tried to actually build the platform. So before I missed out a piece of my story at the start. So I was really fortunate before starting Seal to work at a no code company. So people went a third of them, but they’re doing really well. And so no code is this idea that you could actually build software without writing code. So people might be familiar, SAP is a good example, but there’s also things like Airtable or even something like Notion. This idea that you can end up building systems without writing code. You’re right, yeah,
14:07 – 14:33
Will Moss : and there are places where, if we want to integrate the systems, you do need the technical person. You need the person who knows how to get the CSV out of the RP. But if you have an ERP, you probably have some IT person who can get CSV out of it. But if you’re a small biotech, you probably don’t have an ERP, and then you could just build what you want in our system without writing any code. The other way that we make it easy is that we built AI into the platform. And AI is actually really,
14:33 – 15:03
Will Moss : really good at if you can give it a really discrete task. So if you can say, look, here’s a CSV, I need it transformed into this format, or how do I get this from here to here? It’s really good at solving those sort of little problems. Obviously there are cases that when you just need an IT, you need the expert who knows how that bits of software works. So case by case, with the larger clients, we can provide support services, but we’re trying to put everything as far as possible in a way where we’re just providing
15:03 – 15:06
Will Moss : the tools, everything we try and productize as much as we can.
15:06 – 15:59
Yan Kugel: Right. You mentioned trying to shorten the feedback loops and the updates and everything else. So you gave an example where if a company provides a software and they need to update, they need to go to this company, they do the changes on their end, and then it takes time and then you need to test it. So basically what we’re talking about is creating a more agile environment within the pharma organization for themselves, right? And I think that pharma is, it’s difficult to implement Agile there because everything needs approvals and validation, et cetera, et cetera. So what
15:59 – 16:29
Yan Kugel: is the best way around this? How can you create like an agile environment? Maybe you can say a few words about what agile means for maybe people who are new to this. And what is your vision for an agile environment within the, at least in the systems, systems admin area or the system validation department?
16:30 – 17:12
Will Moss : Exactly, yeah. So for those who aren’t from a software or product background, Agile is a way of working that has been basically completely adopted by the software industry. So anyone who builds software, anyone you know who writes any code at all will be familiar with Agile. And Agile is kind of what it describes. So it’s this idea that at any time, rather than writing out a sort of 12 month plan and saying this is what we’re gonna achieve and these are the steps, you’re taking the smallest steps possible, seeing the results and then adjusting your course.
17:12 – 17:40
Will Moss : And there’s more to it, there’s actually a whole framework around how you build teams, how you plan small units of work, which are called sprints. I think they should be called something else, but that’s a different point. But essentially, you chunk up your work into these very small periods of time called a sprint, which in software tends to be even 1 week. We run 1 week sprints or 2 weeks. And so that means at the start of every 2 weeks, we look, we start with a blank page and we go based on our customers, what should
17:40 – 18:12
Will Moss : we actually build for the next 2 weeks? We try and start every time with a blank bit of paper and adjust what we’re gonna build based on the feedback because we might have ideas about what’s really important for the customer but we’re not the experts. Some of us have backgrounds in quality but we aren’t the ones in the room using our software. They know much better than we do both how our product can be improved and how, you know, what other opportunities there are, so what extra features we can build to solve other problems. And so
18:13 – 18:49
Will Moss : on a very short time frame, we try and sort of reset our vector, reset our path. And so within the the the farmer industry, in many places this is kind of, it’s not, you’re absolutely right, Yan, just not going to be possible to fully adopt in the same way that you can with software. But I think software has shown that making great products, if you can compress that feedback loop, if you can sort of make something, get feedback, make something, get feedback many more times, you’re gonna end up reaching a much better end result than if
18:49 – 19:20
Will Moss : you kind of make all your assumptions at the start and you jump towards the end. And so that applies not just in this case where we’re talking about automating particular workflows or digitizing particular parts of the operations for life science companies. Like it applies even in the actual science itself, so in the scale up, in the cell line development, whatever it is, whatever that particular process is, if you can chunk that up into small things, get the data, I mean scientists are used to this like that, that is their way of working. I think it’s just
19:20 – 19:53
Will Moss : really difficult today that people don’t have those tools. I don’t think we’re going to get in a very short period of time down to those very short people. But there is a future. I can see a future and this is what we’re building towards at Seal. I really believe, and it sounds insane, but at the moment it takes a decade and a billion dollars, you know, obviously it depends on the product, but to bring in like a whole new biologic to market. And for that reason, only 13, I think, came to, they were obviously more biosimilars,
19:53 – 20:26
Will Moss : but biologics came to market last year. And that just means we’re kind of at the tip of iceberg of the potential of the things we can get out there. And there’s a world in which we can cut that cost, cut that time down into days, into weeks. And it’s a useful thought exercise. It’s not gonna happen in the next couple of years for sure. But I like to think, if that was the world we inhabited, what would have changed? And for me, the big, big change is we turn as many of these manual workflows. So anything
20:26 – 20:58
Will Moss : that’s done physically is going to take time. Running an experiment, running a scale up, figuring critical quality attributes, critical process parameters, it’s gonna take, it’s gonna take months and months and months. But there’s a world actually in which we have done this enough times, we’ve got enough, good enough view of the system, everything’s pretty much automated, that probably saves you, in my opinion, like 50 plus percent. So we can get down from 10 years You know we’re down to it we’re down to a number of years, but then we start using the data then we start
20:59 – 21:25
Will Moss : Simulating things without having to do them And then we can cut not just, you know, we can cut years and years off timelines. Because if you felt really confident that given a certain cell line and given a certain product, these were the optimal conditions. This, you could, you know, look at your past data, the AI could tell you exactly what the optimal conditions were, that we can get these things down into days. The hard thing on the other side, this bit’s all easy, everyone kind of knows this, everyone’s talked about it. The hard bit is the
21:25 – 21:57
Will Moss : quality piece. Like how can you get, you can do that, but how can you have absolute confidence that this is the specified product that works the same way and then I mean it’s going to have the right effects of the patient but it’s going to have that those things are really really hard and really really complicated systems but the same ideas apply and I think Certainly in the next number of years, we can reduce the cost. I think we can take multiples of it. Kind of like when you see in the software world, like how fast
21:57 – 22:29
Will Moss : AI is improving or the classic example of how fast, the reduction cost of storage of digital information. It should be an exponential thing. I’m hopeful we’re right at the bottom of that curve at the moment, but we can build confidence and day by day we can go, yes, we’re happy taking bigger and bigger jumps without actually doing the work. And we’ve got better and better protocols for validating and feeling confident that whatever we built is predictive and that the automation is going to work. And we can take step by step. But the first step in that
22:29 – 22:55
Will Moss : whole journey is automation, because without the automation, you don’t have your, it’s very difficult to get all your information in 1 place. It’s very difficult to organize things. So that’s the bit in the puzzle we’re solving. I’m not even sure if we’re going to be the company that solves the other bit, but I see the first bit we can solve is get, automate everything, a unified system of records, so everything’s in 1 place, with the same data ontology, so you know the structure, you know the context. Other very, very clever people can hopefully go and solve
22:55 – 23:03
Will Moss : these other bits, but this is the problem, I think, we’re made to try and help with.
23:03 – 23:07
Yan Kugel: Right, exactly. So concentrate on your strengths, right?
23:07 – 23:08
Will Moss : Yes.
23:10 – 23:12
Yan Kugel: But this is the ideal world.
23:12 – 23:13
Will Moss : The ideal world.
23:14 – 23:46
Yan Kugel: Shorten the time of innovative drugs to the market and automate everything so that you boost your production and you don’t have shortage of medicine. Maybe with that you also reduce the cost of drugs because it’s much quicker and it’s more accessible. So basically drugs cost so much because as you said it takes 1000000000 dollars to bring a product to the market. So if you can cut it by half, you can also cut the price of the drugs by
23:46 – 24:00
Will Moss : half. Exactly. And you can release probably exponentially more. Instead of making 10, if you cut the price in half, probably that ends up with, you know, a hundred new products, not 10. And if you cut it half again, you end up with a thousand. Like it’s not going to be a linear thing. So, right.
24:01 – 24:08
Yan Kugel: Right. Then, the idea is to, you know, to chase the best molecules then, right. So this is
24:08 – 24:08
Will Moss : a, yeah.
24:08 – 24:50
Yan Kugel: Right. And, I think the market will evolve much quicker because right now you have a patent and usually, right now, because you said the patent is valid for 20 years, but as you said, it takes 10 years to bring a product. Until somebody brings a product to the market, they only have 10 years left. Now they need to pump the prices to cover the remaining investments. So if you have so many molecules that are not in patent bridge, but they serve the same problem, they solve the same problem, then you also have more competition on the
24:50 – 24:55
Yan Kugel: market and again it drops the prices. So it’s at the end, everybody wins here.
24:55 – 24:57
Will Moss : That’s a really good point. Yeah, that’s really nice.
24:58 – 25:41
Yan Kugel: Yeah. So hopefully we’ll get there. And to revert back to the idea that we talked about, about the sprints and agile work in pharma. So you said that the usual work is you have maybe a grand design, but you go step by step to achieve it feature by feature. And let’s say a company wants to automate the whole process, but as you said, it can take a lot of time to cover everything, right? So you want to go step by step. So how would you approach it? Would you take each step and then validate it and
25:41 – 26:07
Yan Kugel: already put it in motion? So at least part of the process is automated or would you say, you know, let’s take a year, let’s test, of course, we’ll do the sprints, we will like validate each, maybe, you know, check each step, but then we will validate the whole process. Or do you think it’s also the approach to validate partial processes? What do you think it’s a good approach there?
26:07 – 26:36
Will Moss : I think that’s really good. So I think the latter is what I see most people do. You’re absolutely right. They take up a big thing and they think, at the end, when we’ve got this all right, then we’re going to validate it. I think some of that is because validation is scary, like people are scared of validation. So we put that off, we put that down the road. But what if you were thinking agile mindset, you would think, well, what’s the smallest possible thing I can build here that’s useful? We’re not worried about the most efficient
26:36 – 27:04
Will Moss : path. Yes, if we make all the right decisions, it probably is the most efficient way just to do all the work in 1 go about it at the end. That will be the most time efficient. But we believe we have a lot of humility, and we see that when we do this 1 tiny thing, we’re going to learn something and that might change everything. So the risk inherent in just doing a whole years of work and then finding out a year later is so much that what we should do is do the smallest possible thing. Like
27:04 – 27:36
Will Moss : how can we draw the smallest possible box around our process? So even 1 procedure, you know, doing 1 tiny thing, getting data from 1 place to be, digitizing 1 particular procedure, whatever it is. Like let’s draw the smallest possible box where we see there being a benefit. This kind of the other side of these projects isn’t just illogical. You know, there’s always someone in the room who goes, yeah, this is just logical. If we automate this, it’s gonna save us much because the ROI is gonna be 5, 10X. Like obviously we have to do it, but
27:37 – 28:07
Will Moss : there’s humans involved. Like We’re all unsure. Some of us have more risk tolerance than others. And so the advantage of drawing the smallest possible box and doing that is not just the learnings, which I think would be by far enough, meaning enough. It’s also that you as within the organization, within the culture of the organization, you can say, we’ve done this tiny project, it took us a number of weeks, but you know, we’ve got lots of customers who within even days validation take a little bit longer, but you can get something up and running, working, and
28:07 – 28:36
Will Moss : you can go to the team and go, look, here’s an example of something we’ve automated. It works. It’s better. It’s providing value. And then you’re going to get much more buy-in from the rest of organization to change things Whereas if you kind of try and lump out a big automation project It’s going to be really hard to even get started because you have to persuade people that this thing makes sense that it’s useful that it’s whatever But if there’s a way in which you can do it, which is only a week of effort, it’s very easy
28:36 – 29:03
Will Moss : to get buy-in from your boss, from whoever to say, look, let’s give this a go and let’s try it out. And as a company, we’ve tried to make that possible. So we’re actually releasing a version of our product for free so that people can take it and run it and see if it provides value to them. And if not, we’d love to know why it doesn’t. If it does, wonderful. We’d love to see how else we can help. But we want to make it as easy as possible to learn, to test things in a quality environment.
29:05 – 29:30
Yan Kugel: Right. And do you have experience where companies wanted to automate things and things went terribly wrong? And maybe you have some stories that we can learn from and to plan better how to do things right by examples of things that went wrong.
29:31 – 30:06
Will Moss : Exactly. I guess when I think about that, I kind of see problems at both the end of the spectrum. So a lot of the startups we work with, I mean, some of them, we’ve ended up a couple of times, it didn’t work out well because we’ve engaged with them so early that they don’t know what their processes are. So these are people still doing very early stage development. And so kind of we’re less in a quality world. But There are cases where the same idea is applied. The problem was that they don’t know their process. They
30:06 – 30:35
Will Moss : don’t know what you want. And so if you don’t know what you want, you don’t know what you want to achieve, that thing can change completely. It’s very, very hard to automate. It’s very, very hard to provide value with software if you’re still figuring out what you’re doing and how it works. But if you can describe, you know, where we’ve been successful with those sorts of companies now is going back to what I said before, and we just say, look, you’re not, you don’t know everything you’re going to do yet. But we, there are places where
30:35 – 31:02
Will Moss : you do know what you’re going to do and let’s digitize those. Let’s automate those. This, this make those super streamlined. And then at least there’s benefit we’ve improved. And then as you gain more confidence around the rest of how you work, you can take seal and go and point out those problems. But don’t try and take what they people’s natural instincts. And I think everyone’s is less, they’re so excited that they say, look, let’s go and do everything. And they get 1 guy or woman and say, this is your job, you’re going to automate everything and
31:02 – 31:35
Will Moss : it’s just an impossible task. What you need to do is focus on 1 little issue at a time. The other side of the spectrum is as you get into larger organizations, the problems tend to be more cultural, more people-based. So the battles we’ve had there are more that we’ve got people who love the solution, love the product, huge value, but there’s these budget battles. And so we only end up, you know, I think, so we kind of see it as a steady march. We just do a great job after great job after great job. And eventually,
31:37 – 32:04
Will Moss : people are going to adopt it across the company. But the way bigger companies work is kind of a lot more thought out. Budgets are set. Things are planned. And so it can be sometimes really hard to do a good job because you get budget from 1 team, but they need to interface with this other team who, who are using some other thing and they’re, you know, they’ve got their own budget and their own management and the sort of cultural collisions, particularly when, as you mentioned, right at the start of this conversation, we need to integrate with
32:04 – 32:38
Will Moss : different systems. So we need someone on the other end who, in some cases, who says, yeah, I’m happy this makes sense, let’s integrate it, let’s get that going. The bigger the organization gets, the more chance there is that there are certain functions who are saying, no, this is not how we do things. This is how this works. You can’t integrate with this like this. We’ve learned there’s certain patterns where that happens. So there’s certain teams where they’re not the best place to start, even though instinctively it kind of, we thought it would make sense. And there’s
32:38 – 33:06
Will Moss : certain teams where they’re always really happy. There’s such deep problems that they’re just desperate for automation. And so We try and start with those teams, build up momentum, get people happy, and then we go to these other teams. You can just be like, oh yeah, this is like, at this point they’re like, okay, this is how we do things. And then they adopt it. So, there’s challenges. And what we’re gonna try and do from our side, and we’ve not done a good enough job yet, I mean, we’re still a young company, is we’re trying to make
33:06 – 33:30
Will Moss : everything, as I said, like the product, we want everything to be done by the customer. So we don’t, you know, we want as much as possible that our platform, you don’t even need us, you can just take our product and run with it. And so we’re writing out guides, both the startups and for large companies saying if you want to adopt seal, this is really how we recommend doing it. Like if you’re this sort of company of this size, we think these teams tend to be best or this is how you know whether, you know, this
33:30 – 33:59
Will Moss : is a good way to set up. And this is kind of how to build momentum with your organization, and this is how to get by. And this is how to talk about it even to your seniors and talk about the benefits of it. Because in a bigger company, things do just work a certain way. It’s not something I’m used to, because I’ve only ever worked at startups. And I worked at data scientists that a larger company wants, but I hate to hate it. So, but it’s how the world works. And we have to give people the
33:59 – 34:04
Will Moss : tools so that they can navigate these sort of more difficult situations in the best way possible.
34:05 – 34:36
Yan Kugel: Right. And what would you say are the biggest scares of companies to embrace automation and AI and external tools. So the biggest scares, which are actually have very easy solutions. We should say, you know what, I’ve seen so many companies and this is their biggest scare, but then it’s nothing. Good question.
34:39 – 35:23
Will Moss : So I think it’s 1 I can really relate to. And it makes complete sense. If you’re a company making a life changing product, you’ve got all of your IP, you’re the experts, you want control over everything, you don’t want anything to go wrong. Like bringing in a software or system that you don’t feel in control of is really like, of course, it’s terrifying. And so I guess that’s what we’re trying to do is say, look, this is your tool, your software, take it and run with it, however it works for you. That’s the solution there. The
35:23 – 35:54
Will Moss : 1 we can’t get over, and it’s just the nature of the product, is there are still people in the industry who are scared of the cloud. And that’s, It was surprising to me when I first got into this world 6 years ago, because everywhere I come from, it’s taken for granted. And I thought the world had taken it for granted, given where we use our phones, the way we use our laptops at home. But there’s so many companies. And I even speak to consultants and quality people who are experts who he does. I disagree with really
35:54 – 36:28
Will Moss : strongly and he disagree with me and I’m sure this case is where they’re right. But they really do not believe things should be in the cloud. And they have a good sort of list of worries, which do make sense. Like, what if it goes offline? Like, we aren’t in control of it. Someone else could be hosting it. That’s not a silly tree, But where is our data? Do we have control over it? We’re not in control of the updates necessarily. Someone else can come do them. There’s this whole very good, sensible list of worries about the
36:28 – 37:02
Will Moss : cloud. But We’ve seen in every other industry, even like super regulated industries, you know, finance, everything, everything is in the cloud and there’s ways to build things in the cloud that are really secure, that are really thought out, that are really sort of carefully set up. And the counter that I try to say is like, well, let’s think about the inherent, you know, whenever you’re making a change, I think you should try and think about both the risk of the change because that’s what’s present. Like the change is the scary thing. But what you should try
37:02 – 37:33
Will Moss : and think about what I try and think about in my life in whatever it’s like, there’s also inherent risk in not making the change. There’s inherent risk in the way I’m doing things today. So like, you know, in your life, You could be like, I don’t like my job. I’m scared of quitting. I’m scared of whatever. And that’s obviously risky and scary, but that’s the risk you feel. You feel the risk of not having a job. You don’t feel the risk of being dissatisfied with your life and not being happy and not being supported. So it’s
37:33 – 37:58
Will Moss : the same with the cloud. But there’s such inherent risk of having things on a server every like, where’s the data? How is that secure? Like this, we can reel them off. It’s we’re using technology from 1990. Like everyone, We can’t use the best tools out there. There’s just so many problems with it. So that’s the way I try and phrase these things, but the situations where a company just isn’t ready for the cloud and we can’t help it.
37:59 – 38:47
Yan Kugel: Right. So as you mentioned, people are afraid of the cloud. They’re afraid of the data privacy, what’s going to happen with their files and their data. And I guess it’s very secure right now. There are a lot of secure solutions and protocols, right? I think this is something that companies shouldn’t be afraid of. The company that provides a service establishes some specific protocols or works with specific cloud solutions. So let’s say companies have a fright of that. So what tell them to check with the service provider, a couple of keywords that if they know, okay, they
38:47 – 38:55
Yan Kugel: are following these rules, they’re using these intraship protocols, you know, you’re in the safe. Is there something like this? Yeah.
38:55 – 39:30
Will Moss : So every software company you engage with should have really good documentation around how your data is stored, how it’s backed up, how their system is compliant, how to use it in a way that is compliant, how you can use it in a way that isn’t compliant. And most companies have that, but what you really, really want, like what would give you total confidence is this idea that you have access to all the information. You can extract it if you want to, and you can even set up systems to continuously pull the data out so that, 1,
39:30 – 39:59
Will Moss : if you’re not happy with the product, you can go and get the data out and go and do something else. You can load it all into your own database and just query it or do what you want with it. And so we do our best on the document citation side of things, but again, like our philosophy is always like, this is a problem, how can we solve it with software, with a product perspective? And so what you really want from your provider, level 1 is like they have a full API and so you can extract your
39:59 – 40:28
Will Moss : information. Like There’s a way in which you could set up, and it’s not hard to do, set up code that goes and pulls that data out and you can sync it somewhere. Ideally, they didn’t do it for you, but if I was you, I’d probably do it yourself. And so you know you’ve got your data backed up somewhere, you’ve got it in control, no matter what happens, you feel in control. Level, a level above that is that you can actually, while using the system, see how the data is stored. So they’re public about their schema. They’re
40:28 – 40:55
Will Moss : public about how all the data is organized. They’re public about, and they’re open in their philosophy about, and they see it as your data, not as their data. And so if that’s the case, then you can, when you kind of have those logs of the data, you can see exactly if things have changed, you can set up your own. I think the best thing to do is to be in a situation where you actually can take control and not just throw full trust to them unless you do trust them. Like as far as possible, you want
40:55 – 41:20
Will Moss : a provider who is gonna make you feel in control of your data and your process. And so, yeah, 1 documentation, 2 really push for a full API or that they’re going to give you the backups automatically. And then 3, yeah, the kind of the gold tier would be that you know exactly how your information is stored. You know exactly how the system works. And like you can sink that somewhere where it’s mirrored.
41:22 – 42:00
Yan Kugel: Great tips, well, great tips. So we covered so much and it’s so, the topic is so fascinating and it will create a revolution in the industry at the end because we see a lot of companies, innovative as seal, and it’s amazing to see that. And I would love to continue this conversation for hours with you. Unfortunately, our time is coming to the end. And I wanted to ask if there is something that you would like to share that you feel that we have it addressed, that is very important for people to know or anything that some
42:00 – 42:02
Yan Kugel: insights that you want to add?
42:03 – 42:33
Will Moss : I think I’d more take it as an opportunity for me to learn. So obviously you’ve got a wonderful audience of smart people who know lots of things that I don’t. And so I might be wrong, you know, I’m out to some of these questions. There might be the reason that I am wrong. Give me the I would love to know the reason why you ask it. I’d love to know the reason why it doesn’t make sense to get an agile way or like to help to help me up or you might just say actually will there’s
42:33 – 42:59
Will Moss : these other things that maybe you should consider and think about, that’s what I would love. So if, and I guess that’s kind of me trying to fully embrace that agile mindset is that if you have thoughts of this conversation, I would love to hear, I would love to know whether you agree, whether you disagree, whether there’s extra thoughts, please reach out, please tell me and I’d love to chat and I’d love to get your advice because I’ve not been a quality professional in the way that you have and so there’s things that I don’t know and
42:59 – 43:02
Will Moss : so please please reach out and please tell me.
43:03 – 43:22
Yan Kugel: Nice. Thanks, Will. So we’ll have Will’s LinkedIn profile in the description of the podcast, also the email address if Will wants to share it. And feel free to get in touch with Will. Check out the Seal website. So it says seal.run. Yes. Very interesting.
43:22 – 43:26
Will Moss : Yeah, and I use the domain, but we like the idea of run as in running a process.
43:26 – 43:36
Yan Kugel: Yeah. So very cool. Very memorable. So seal.run. Check it out. And Will, thank you very much for this great chat.
43:36 – 43:37
Will Moss : Thank you, Yan.
Yan Kugel is joined by Christy Mazzarisi, a distinguished leader in the pharmaceutical industry with extensive experience in quality systems and leadership roles. In this episode, we explore Christy’s journey […]
