Arm Viewpoints: How AI and biometrics are powering equitable immunization in Ghana

Brian: [00:00:00] So let’s start by having everybody just go around, introduce yourselves. We’ll start with Tristram Norman, co-founder and chief technology officer at Simprints.

Tristram: Hi guys. I’m Tristram. I’m as you were saying the chief technology officer and one of the co-founders of Simprints. I started Simprints 10 years ago, almost 11 years ago with the other co-founders.

I have a background in technology and I just love solving really difficult challenges. So it’s a pleasure to be here.

Brian: And Bertrand Pedersen, senior Manager, private sector partnerships and innovation with Gabi, the Vaccine Alliance. Tell us a little bit about yourself and a little bit about the Alliance as well.

Bertrand: Hi everyone. I’m part of Gavi, the Vaccine Alliance. I’m working on our private sector engagements and our innovation efforts at Gavi, needing the in infuse platform. Gavi is an alliance that is aiming at vaccinating children around the world in low and middle income countries leveraging the expertise and the innovation that [00:01:00] our private sector partners can also achieve.

Brian: And last but certainly not least, Rosalie from. Arm, Rosalie Trod. Tell us a little bit about yourself and your role at arm.

Rosalie: Yeah. Hi everyone. I’m Rosalie and I lead our social impact and innovation partnerships here at arm. So my role sits at the intersection of technology and global challenges where we’re working with social enterprises and not-for-profits, such as our partnerships with St.

Prince and Gabby, the Vaccine Alliance to explore how advanced technologies including AI, can really strengthen real world programming.

Brian: I’m gonna make this very convenient. I’m gonna throw to you, Rosalie, you all were just in Ghana working on a fantastic project there. Talk to us a little bit about that.

Set the stage there, and then tell us about the identification gap. How do you define that? What does that mean in the context of the technology and the services that we’re talking about here?

Rosalie: Have been working together on, on specific programming in Ghana, rolling out [00:02:00] Prince technology there. And I think really as you said to, to set the scene when we, it means that health systems often don’t know which child received which vaccine, and that uncertainty can directly cause missed doses, duplicate records, children left unprotected.

And when we say identification gap, it can sound a little abstract, but globally there are about 850 million people that don’t have access to formal id. Many children are effectively invisible to health systems. And I think on the ground and what we saw in Ghana that can show up as fuzzy data where official coverage numbers may look high.

Yet local clinics are reporting impossible figures. Some districts sometimes recording over a hundred percent coverage, and often that’s ’cause of duplicate fragmented records. And it’s not. It’s not really a statistical quirk, it’s often a real failure to link people to their health history because they don’t have that unique identity.

I think that’s really where Simprints’s technology is really coming into play to provide people with that unique [00:03:00]identifying biometric record.

Brian: A big challenge, obviously and also a big opportunity. Ra, tell us a little bit about why. Multi-dose vaccines are important and those are vaccines like the malaria vaccine.

What is it about this gap and having to deliver these vaccines? That’s a challenge for the Alliant.

Bertrand: First of all, I just really want to mention that we are in a golden age of vaccines. 2024 was really a historic year with the first launch of the malaria vaccines. We also have a TB vaccines, which is the world’s deadliest infectious disease killer.

On the horizon, we have mRNA technologies unlocking endless of future breakthrough. And this progress also puts us in a position where we have to optimize equity. Access to ensure these vaccines really reach every children we need them and have the maximum global impacts. And so with the Aria malaria vaccines it’s a great [00:04:00] example.

It’s a multi-dose vaccines, but it’s also the vaccines that was rolled out the fastest than any other routine vaccines in Gavi history. So we are now present with malaria vaccines in 24 countries. And two different vaccines that are both proven, safe, effective, and really set to slash children’s death.

However, it’s for those vaccines it takes it starts around five months off. Five months old with a fifth dose a year later. That makes sense. In high seasonal or persistent risk areas to extend the protection. So both of these vaccines they kept malaria cases by over 50% in year one, and that’s when the kids actually face the peak risk of illness and deaths.

And the fourth dose is boost longer term defense. So those multi schedules are. Hugely important to make sure we have the maximum impacts, but also they’re like drop off magnets. Each extra clinic visits. That’s an additional [00:05:00]barriers, travel costs, caregiver time clean to access any daily priorities that brings more failure points and having those.

Weak records also kills those completion. It’s not clear who’s due when that increase the number of miss kits that we also have. And that’s why having this gap is particularly a challenge for us.

Brian: So you are all. Just in Ghana doing the actual work there. Give us a sense for what you saw, what you experienced there.

I want to, I wanna hear from Tristram this time. You are the technology that bridges the gap.

Tristram: Yeah, so we saw two projects when we’re in Ghana that St. Prince is running at the moment. So the first project, which we call GG two or GAVIGAN two, its goal is to digitize the community health workers workflows at the chips facilities.

So most of the healthcare in Ghana and many regions around Africa like that are done by community health workers who are amazing people, but they have to go to [00:06:00] far away rural places and deliver care or aid to these different places remotely, essentially. So they work in small clinics. So the community health posts is what they’re called.

We call ’em chips. Facilities are usually small clinics in a rural Village where they’ll treat people and then they’ll refer them to one of the larger hospitals. Our goal is to digitize the workflow. When we were there, we talked to a bunch of Guinea health workers. Really great one, Mary. She showed us.

How they used to have these paper-based records. So we went in the room, we saw this massive book that they used to have to carry around with them when they’re doing their visits, and they would try and write down mother’s names, children’s names, and as Bertrand was really talking about, imagine taking this book around and trying to do a four or a five dose vaccine rollout.

If you can’t find the previous visits it becomes a really impossible task. You don’t know, like, how often have I seen these kids? What dose are they on? When do I follow up? Who should I be [00:07:00] following up with in this area that I’m visiting? So our goal for that was to work on digitizing the entire health workflow, and specifically starting with vaccinations but also supplements as well.

And so we work with the Ghana Health Services and they we help set them up with this platform so we can digitize everything. And then sim print’s goal in there is, can we use biometrics to reliably. Find these records every time that we visit them to pull up the right record to make sure that we’re giving them the right vaccine, the right treatment.

The other project that we saw is we call it a IB, it stands for ai. So for AI infant biometrics, one of the biggest biometric challenges is infants. So when you think of biometrics and you’re matching your face to a previous photo of your face that you have, you’re logging into your phone, these sorts of things.

It’s really designed and built for adults. So it’s very much doing a one-to-one match of. Does this person’s face look like the face I’m expecting it to look like, [00:08:00] and we have a couple challenges here. One, in our situation, we actually have to look at it and say, where out of these 10,000 potential faces can I find this person?

And then the biggest challenge is doing this for infants. ’cause imagine how much an infant face changes across six months. Yeah. They look very different even after just six months. So our goal and what we would love to do is to try and bring the age where biometrics for LEB works all the way down to day one.

Can we register a baby at birth and then when they come in for their vaccination, still reliably find that baby. So we’re doing a big project where we’re collecting a lot of data out in Ghana at the moment. All for infants between zero and 12 months old. And it’s a huge longitudinal project where across a year we’ll see an infant and their mother 12 times or their caregiver 12 times.

And we’re hoping to use that to generate an algorithm that will be able to accurately do this so we can [00:09:00] continue doing exciting projects like Gagan too. Basically,

Brian: obviously you’re doing some pioneering work here. Just tell us a little bit about how. Simprints and Gavi connected who reached out to whom or originally to kickstart this?

Tristram: I’m sure we reached out to Gavi. In this case, I think we’ve worked with Gavi on a few things and especially because our use case in biometrics and what Gavi needs to do, multi-dose vaccines lines up so well. But it was Gavi who really came to us and said, look, this is exciting.

They love that SIM prints. Our focus has always been not just like biometrics and the cutting edge, but really biometrics in these settings, right? How do we get biometrics all the way out to the edge, to the really last mile users here? They said, and that’s great and we love that you guys are doing that, but you need to do it for babies.

You need to bring this technology and make it work from day zero, which doesn’t exist at the moment. And so we can start treating babies with the malaria vaccines and all [00:10:00] the other exciting things that they’re working on. So we were excited for that challenge. There’s a long way to go still.

But yeah, we love the work so far.

Bertrand: And if I may add this whole Gavi Simprints relationship, stem, stem from the focus that Gavi has on innovation, it’s really been part of our DNA since our launch. And actually in 2016, Gavi Launch Infuse, which is a mechanism, a platform to identify innovation that can help us in our mission.

And that’s how we launch a code for application where sim prints applied and we identified them as having huge potential to to help us with that. And Ghana is a perfect breeding ground of to try those technology. We’re not just importing technology in the country here. We’re seeing Gavi reported countries actually leading global breakthrough working with partners on this first AI enabled, infant biometric is a pretty extraordinary proof point of that, so it’s amazing to bring Gavi Simran together and in a country where we operate.

Brian: Rosalie, talk to us a little [00:11:00] bit about what you might call sort of the consumer experience, right? Is bringing the service imprints, bringing the technology, but then you’re addressing a demographic. It may be unfamiliar with this sort of technology, but we’re making it really easy to use. Talk a little bit about that.

Rosalie: I think being in Ghana a few weeks ago really brought that to life in such a human way. And I think what’s, what struck me and Tristan’s mentioned this slightly too, is, how central the community health workers are. Like, they’re the system, they’re the healthcare system. They’re that first line for people in Ghana, they know every family.

Like they’re juggling all of these paper records, immunization schedules. They’re doing it in clinics where power and connectivity are not always guaranteed. So I think any technology really has to start with. And with Sim and Gavi, I think the focus has really been on making tools that are genuinely usable for.

As you call it, the consumer or for these [00:12:00] community healthcare workers. And it’s a simple app that’s on an arm powered smartphone that can work offline can give real time feedback and will fit into those existing workflows of the community health workers rather than fighting against them.

And I think the other thing that we saw when we’re in Ghana is the importance of the community sensitization. So that’s why the caregivers are really the community health workers are really explaining to caregivers like what biometric ideas, how their data is protected and why it’s really helping them to make sure that their child.

Isn’t from healthcare. And for me, I think that’s really the heart of this work is that user experience, and it’s not just about this new impressive AI algorithm that we’re looking to create. It’s about how simran are really co-designing that with the health workers, with the communities.

So that solution is actually really working in the clinic for those health workers, not just in a slide deck. Tristan mentioned one of these particular community healthcare workers that we met, and she was [00:13:00] very enthused while she’s still working on both digital and paper-based records was very enthused for the moment in time where the digital records and the Simprints technology can take over and they can get rid of the paper-based records.

And she really saw how that really helped her work and really sped up her work and the accuracy of her work.

Brian: Tricia, let’s roll up our sleeves a little bit on, on technology here. Biometric technology is amazing in the lab, but you’re making it work in real world conditions, limited connectivity, power, battery issues, harsh.

Environmental conditions. Talk to us about how you make that happen.

Tristram: Yeah, I think it’s a great question and it’s one of the reasons that Simprints and Arm have always worked so closely together and why we’ve always really loved working with arm because it’s so important to the work that, that we are doing in the field.

So Simprints, even though people hear the name and they think biometric company. And so many biometric companies are really focused on just cutting edge biometrics. How do [00:14:00] we make, identity security, these sorts of things. Simprints, we focus on cutting edge biometrics, but so much of our work goes into how do we make that work in the last mile?

How do we make that work? In places that have limited to no connectivity? They might be offline for weeks to months at a time. They have very low powered devices. They have a really hard time getting Android devices in the first place and sourcing them, and they could be, a decade old. Some of these devices.

And so it’s about how do you take these powerful things and these algorithms that we do have and they are available in the Western world, they do work, and how do we make them run in that sort of setting? And I think everything in our stack at some point relies on arm technology, right? The core one for us is the devices that we have in the field, but I think something to try and really drive home how important.

Having those be performant and designed for the edges. If you think of, you write a [00:15:00] message to open AI or philanthropic, whoever you’re using, right? And you make this request and it goes off to the cloud as we call it, and there’s some giant computer and there’s some big servers somewhere and this massive model is running and it’s computing all these things back and it sends you back an answer.

And for us, in this Western context, that’s great. It works really well. It works perfectly. We have great connectivity. But think of a challenge that requires something is with a heavy lift like that, doing these biometric comparisons, these computations that we need, but we don’t have any connectivity.

This needs to work on a hundred dollars Android device. And not just that, but it needs to work reliably. It needs to not completely flatten the battery of the device after a couple hours of use. It needs to be able to go out to the field all day and needs to run and work in the dust and the heat. And that’s a really important thing of what we’ve always been focusing on is.

How do we make it run on the edge and arm chips are what’s powering our ability to do that. We can make models run efficiently on them. We can do [00:16:00] this amount of computation on those low end devices. It’s really important to, to what we do. That’s always been our focus, is how do we help people like the community health worker that Rosa was mentioning.

Mary, her day is, she’ll be at the clinic all day. If this device and this workflow don’t work, that’s a massive headache for them. They, when we go in these clinics and it’s great and they work so hard, but there’s 20, 30 mothers in there, they all have babies. They’re all trying to get somewhere.

They’re all trying to receive a service and if they’re spending 20 minutes on each mother, ’cause they can’t get the signal to connect or the need to wait for the phone to charge, or even if just it’s slow. It’s hard for us to imagine, but it just becomes such a difficult, situation for them.

So making that smooth as possible and reliable as possible is really the thing that we’ve been well spending the last 10 years trying to focus on, trying to do.

Brian: And you mentioned earlier of the challenge with infant identification. That phase changes very rapidly, so on [00:17:00] and so forth. How are you using AI to.

Improve that identification.

Tristram: So the goal here, right now, we’re starting with data collection. So when you build a model and when you try and design an algorithm for this use case, you need quite a lot of data, right? And as we know these AI models and a lot of machine learning is done, has huge pools of data.

So our first task is to collect a huge amount of data we’re trying to collect. It’ll be after a year, it’ll be millions of images, something like 16, 20 terabytes of data that we’re trying to get outta the field. In terms of images, it’s a lot to sync up. And the key one that we’re really starting with in terms of AI actually is.

Collecting biometric data, the biometric samples is very difficult. It sounds like you’re just taking a photo of a face, but it’s not, these faces need to be accurately aligned. We’re looking at photos of ears. So we also need to know is the [00:18:00]ear, is it blurry? Is the contrast right? Is there good enough lighting that we can see all of the key components that we have?

Is it oriented in terms of is its rotation correct? Can we see the ear flat on seeing the feet? Imagine taking a photo of a baby’s foot. It’s moving, it’s kicking. Hopefully many of them are sleeping, which is really great. In, in real world use cases, it’s difficult to do to collect this data.

And it’s not just that, but it’s the community health workers collecting this data. So we need to give them real time feedback. Now, if we, if they take a photo of baby’s face and we just say, the photo’s not good enough, do it again. Imagine if they have to do that four or five times, that’s, it’s a very unbearable experience.

It’s not something that you want them to have to go through. So the first thing that we did with AI is we went out and we collected a lot of small sample sets, and we trained a couple models that were really good at giving feedback of, is this photo that you’ve taken of a baby’s face or their hands, or their ears or their feet, is it good quality?

Is it the type quality that will need to create this algorithm? Hopefully that can work in the future. And now [00:19:00]that we’re going through this big data collection, we’re about month four, I think. Month three and a half, four-ish, and we have about 2 million images. Now it’s about assessing the quality of those images, making sure that it’s hitting all the markers and the metrics that we need.

And then after that, we’re hoping to use those images to take, there’s some larger foundational models out there used for biometrics, and we’re gonna be exploring can we use that? Can we retrain them, fine tune them? Can we take some of the layers and freeze them and make them work down for infants? We’re also exploring, can we look at aging, synthetic aging?

Can we say. You capture a baby at day zero and then come back two months later for a followup vaccination or some other delivery. Could we actually say we have the original capture of the baby? Synthetically age them two months and then compare that to the baby who’s two months old sitting in front of me.

Or do they look similar to what we’re expecting and use that for matching. So there’s a few things, but we have, there’s still a lot of work left to do in terms of getting the full distance, in terms of getting accuracy that we need.

Rosalie: And [00:20:00] can I jump in with a sort of observation from when we were in Ghana, Simprints got us all to test what it’s like to try and collect this data, which was a really fascinating thing to try and do.

And surprisingly harder than we expected and really helped us to understand. What some of the challenges are for the community health workers that are collecting this data. And I think it’s super important that Simprints is, they’re trying that too, because then they can really iterate on the product to support the community health workers.

But what I found really fascinating is obviously Simprints are starting with, I think you started with four different unique biometric modalities initially, which has faced here, ear palm print, and heel. And what’s really fascinating and a little fact that I loved is that actually palm print’s been ruled out because.

Trying to collect a clear image of a baby’s open palm is really not practical at all. And often they’ll be asleep and have their fist clenched and trying to pry that open will upset them. Really interesting to see that. Just something really practical [00:21:00] that has ruled out being able to collect sort of palm prints.

Still collecting ear and hill prints was surprisingly challenging for those of us trying it for the first time. And it’s really impressive how the community health workers have. On board and become experts at doing this.

Brian: Bergstram from Avi’s perspective, why is digital identity, which you’re now capturing right there with this great technology in the field, why is this becoming foundational to equitable immunization programs around the world?

Bertrand: We see really how everyday, how technologies is reshaping public health. We, it brings new opportunities, it brings also new responsibilities to us. So digital identity, ai. Machine learning, drone delivery, they’re all changing how country can deliver monetary immunization programs and digital identities is so foundational for Gavi because simply because it ends the global identity crisis.

We have millions of children, especially in remote poor areas that [00:22:00] lack any formal record blocking, accurate tracking and deliveries of vaccines. And so we see digital identity as a key to re to registering, verifying those kids from birth, but also pulling zero dose children. So zero dose being the children have never received any dose of vaccines as well as under immunized Children’s into the system for a more targeted reach.

And we’re talking in depth today about the Simrin solution, the ARM support. And I want also to give another example of how technology is supporting Gavi. Also with the great where we are grateful for arms support is our partnerships with Atlas AI and other infused innovators from our latest cohort of innovators.

We’re collaborating together in India and we use geospatial platform that, that leverage artificial intelligence and data analytics to identify and geo target. In the last mile to find those zero, those children and miss communities. And more specifically with [00:23:00] these partnerships, we, we allow to to map community health facilities, immunization program the coverage of this immunization program and integrate village level vulnerability and outbreak risk.

And we analyze how those risk factors create the right condition for potential. Adverse impacts and outbreaks.

Brian: Amazing work. Let’s go back to the recent trip and engagement you had in Ghana. You get a feedback loop when you’re there on the ground. Rosalie, talk to us a little bit about what you’re hearing from a.

Frontline workers and then everybody else can jump in if you want. How you’re taking this feedback and iterating on the process and on the technology.

Rosalie: Yeah, so very happy to share what I heard from the health workers and what we’ve heard from the families. And I might allow Tristan to then talk about how sim prints are integrating that feedback live via WhatsApp with the community health workers chasing them, I think 20 minutes if they haven’t responded.

But [00:24:00] one of the most encouraging things I think that we had from the community health workers is that the system is already making their work faster, making their work more reliable. And I think actually even in, in other countries, Simrin has seen community health workers being able to identify clients around three times faster with biometrics and to reach record accuracy levels compared to the paper-based systems.

And we are seeing these similar patterns emerging here as the data comes in Ghana. And from a user perspective, I think they really value the confidence that it gives them because they’re not. Whether it’s the right child, they’re not having to flip through the giant, like a one huge heavy paper records to try and find out.

Which, which record it is that they need for that particular child. But at the same time they were honest about some of their pain points. And right now many of them are still running both paper and digital and parallel, which is extra work. So I think that’s something they’re really keen to move beyond.

And I think particularly for the families and especially the sort of caregivers. It’s great for them to know that even if they might have lost their child’s vaccination record or ID [00:25:00] card, or if the child is brought in by a different caregiver, there is always gonna be a secure way to find that child’s specific record and connect that child with their record without that paper.

Paper based id and I think the combination and the continuous feedback from health workers and the trust that it’s building with the families is what allows imprints Gabby and Arm to keep being able to iterate and improve on the product together. And Tristan, can I hand over to you to build on how you work with the community health workers?

Tristram: Yeah. I think, we’ve always believed that you have to be user-centric when you’re rolling these things out. We have so many great examples as well as tension to simran and Gavi and Arm can be if what you’re building isn’t working for the user, then it’s not gonna have the impact.

It’s not gonna help. And there’s so many great examples of great technologies that do work, but they failed because they weren’t centered around what the user needed. So we’ve always tried to focus on. First, what is the real need of the user? And then second, what is the technology?

How can the technology help that need? And I think that’s the big one, especially for Gavigan [00:26:00] two, has been about starting smaller. We started with a few clinics and a few chip facilities getting very close and working very closely with the community health workers. Seeing how amazing and motivated they are, but making sure that we’re also understanding what does their day actually look like.

It’s not just, is this technology coming? Is it sexy? Is it fun, but is it helping? Is it saving time? And then measuring that. So we were looking initially at a. Can we improve how the chips outreach and the follow ups are? And we’ve seen in the small clinics 20% already improvement in terms of outreach.

If they get quick notification saying, Hey, this person was supposed to come in today and they haven’t, can they act on that much faster than they can at least paper based? And a big one, a really big one, is eliminating duplicate records. It’s very easy in paper-based systems and even in electronic medical records systems to create duplicate records, and then it becomes much harder to find the correct person.

So it’s like taking that feedback and then iterating off of it. And then, so we [00:27:00] took that and we moved up to 50 clinics and now with the goal of 580 or so wide clinics, and we’d love to one day try and reach, national scale, but there’s thousands of clinics. There’s a long way to go to, we get to that point and it’s a multi-step journey.

It’s not just write code, create cool looking app, give it to users that works. It’s about. It’s about training the users. It’s upskilling the community health workers. We already see how hardworking and how dedicated they are, but they’re not always entirely familiar with the Android devices that they’re working on.

They might not have been able to use devices like this. They’re not familiar with the types of record keeping systems that they have. So how do we upskill them? How do we make this beneficial to what they’re doing? Can we get more support? As Rosie was mentioning, one of our key things is making sure that we have people there who are working in the context with them day to day, and then they feed back to us like, Hey, I’m on this device, this community health workers [00:28:00] having issues.

It was actually Christmas day for me. And one of our, one of our collectors in the field who works for Simprints is really amazing. But he is saying, I’m sitting here with three community health workers. I need you to update these records and it 20 minutes before responding, and he is going, are you there?

I was like yes. I’m here. I’m here. I’ll update this for you. But they’re very passionate and they want it to work. So it’s it’s our responsibility to make sure that we’re taking all the feedback that we can get and iterating on it as much as possible.

Brian: So you mentioned.

A long journey. Let’s wrap up here and I’ll throw this to you first, Tristan, what does success look like in the short term, 12 to 24 months, let’s say. What are you learning from the partnership and how do you intend to apply this going forward? And then we’ll get the last word from Bertran from the healthcare perspective.

Tristram: Yeah, it’s great. I think from our perspective. Success for the two projects for Gavigan. Two success really looks at, can we keep scaling this up? Can we [00:29:00] keep digitizing their record system? We’ve seen big benefits of it. We would love to see this reach wider in terms of the community and further up and the Ghana health services chain.

And so we really about, can we take what we’ve done for a couple clinics and that 50 clinics and bring that to 500 clinics, bring that to 5,000 clinics because we see that it’s helping and the community health workers are excited about it and it can have a big impact. So we’d love to do that. And for a IB, this is a research project.

This has never been done before. It’s a multi-year research project. We’re doing the data collection. Now we see promising things coming outta that data collection. But for us to really. Reach that point where we think that it’s gonna reliably work and where we can help Gabby is can we actually get in from day zero and ID them when they come back six months later?

That success would be can we do that? But to get to that point, there’s still years of research left to do once we’ve [00:30:00] collected all the data and we have very promising indicators that we’re heading in the right direction. I also think we have quite a long way to go and hopefully we get there, but we’ll see.

Brian: Bertran talk to us a little bit about how you see the future evolving. ’cause this is just an amazing human-centric space.

Bertrand: Yeah, absolutely. I think for me the future will be based on two lessons that, that we see with this partnership. So one of the thing that really give me the greatest reason for optimism is this pace of innovation.

I’ve mentioned the vaccines that we’re developing faster than ever before. We also have all these amazing technologies emerging like biometrics and AI and Joe that I mentioned as well. And this solution brought together means that we can forecast where vaccines are needed with greater accuracy than ever before.

We we can get those vaccines to even the most remote communities and we can monitor progress with those digital tools. And this is thanks to innovators like Sim [00:31:00] Prince who bring those crazy ideas into actual concrete solution on the ground. Thanks to founders like I, who willing to take those risks on those ideas to, to the political leadership that we’ve seen, for example, in Ghana where they’re accepted to, to test and help improve those technologies in their countries.

So this is my first lesson is the power of collaboration. This blend of. Public health work, cutting and technology coming together. And another one that I really want to emphasize, and that goes beyond just the identity or trust question. It’s for all the transformative potential that these technologies have, their adoption, their implementation can still be.

Very painfully slow. And too often we see those technologies that have real promise in, in pilots and trials, and they never really reach any type of scale, national or regional scales that are required to really have large scale and sustainable impacts. And that’s heartbreaking because it means millions of people are missing out on vaccines, on, on [00:32:00]health, on any support that this technology could bring within their reach.

And it means inefficiencies, it means squandered potential. What if we could scale same prince nationally in Ghana? What if we could scale to new countries? If I take the example of drone delivery, we’re now present in five countries with Gavi and our partners to, to deliver vaccines. But what about the rest of the continents?

So one of the key challenges that we have ahead is. To translate that innovation into impacts, and that means reaching more scale, more sustainability. And meeting this challenge goes really to the heart of G’s new strategy and the upcoming scale up facility, which is a platform that intend to solve those challenges that we’re currently developing.

Brian: I, I have to say, it’s hard not to get choked up listening to this story, right? We do so many, so much technology, storytelling, and the applications are. Not nearly as concrete as what you’re doing in Ghana, so to you all on the work that [00:33:00] you’ve done so far. And best of luck going forward. Thanks for your time.

It’s been great.

Tristram: Thank you. Thank you for having us.