So I’m here with Roger Noble and Steve Taylor, co-founders and now CEO and CSO respectively of Zegami Limited, a spinout of the University of Oxford. Zegami is a software platform for exploring image data, which combines the images with other metadata to provide a really user-friendly way of analysing complex visual datasets.
Could you tell us a little bit about yourself?
RN: Sure, my background: originally I studied computer science and multimedia, and for a lot of years I worked as a software engineer developing various different solutions for a lot of people back when I was in Australia. Since then I’ve moved into the data science side of things -it’s an area of big interest to me.
It’s about giving people a better access to data and making data more available and accessible to people. And that led me to start to work on Zegami – a tool for data accessibility and data visualisation, and then through that work eventually I got in contact with Steve Taylor at the University and that began the foundations of Zegami. We ended up spinning it out as a company in 2016.
And what’s your background?
ST: My background is: I’m a microbiologist by trade, but I was always massively interested in computers from an early age- I’ve been programming computers since the 80ies- and I love microscopy, -and so combining images and data, and biology is like a dream come true for me. I’ve worked in various different companies from GlaxoSmithKline. I did a start up in Australia – a bioinformatic startup. I worked at a small company. doing proteomics, lots of glycosciences. I joined academia in about 2013 to set up a big high-performance supercomputer and you get lots of users to use by bioinformatic software. My big goal really has always been to get people to use their software really efficiently and make it as usable as possible because that’s where we get the most insight.
Could you tell us more about Zegami, and what the problem is that you are trying to solve?
ST: In biology, we have often have huge amounts of data and we have increasingly huge amounts of images around patient data, cellular data- all sorts of different areas. So, let’s take, for example, ultrasound. We have a picture of how the blood flow works in the heart. We might know the age, the weight, the lifestyle of a person, and we may be able to predict that person is going to get the disease.
By using techniques and AI, we can look at thousands of features around the ultrasound image, some of which are useful, some of which aren’t, and then Zegami would allow you to combine all this data and look at trends across different patients, across large amounts of ultrasound data, and show what’s useful and visually show these trends. It’s also really good at spotting outliers, for example, and looking for problems in your data. It’s got a whole range of use in those areas, and it puts it in the hands of scientists and really democratises the way we can understand how machine learning and AI works in the context of images.
How big do you believe the opportunity could be for Zegami both in the UK and globally?
RN: We see a really big opportunity for Zegami, especially in the scientific and medical research side of things, which is our big focus at the moment. I think it’s pretty fair to say that it’s easier and easier these days to generate more and more data as these sort of tools become more prevalent. But the issue is how to deal and manage all of this data that is becoming more available, especially on the clinical side of things, where these people are typically really time poor and the burden that they have to produce results faster is only ever increasing.
Tools like Zegami that can allow better insight into these machine learning models enables them to, not only be more efficient with their time but also give them better insight into that black box machine learning model, and give them explainability, so they can produce better patient outcomes.
Who are your customers and what’s the revenue model for the business?
RN: So we have a really diverse range of customers actually. We have a data tool and you can put any kind of data into it. It means that we end up with all sorts of different use cases. One of the really early ones that we worked quite closely with was in plant phenotyping. And this is helping researchers to develop better, more resilient strains of crop for the agricultural industry, effectively helping them to grow better crops to feed the world.
But actually, since then, we’ve focused now more on the medical side of things, and we have customers not only here in Oxford, but in Australia and the US and in Europe as well. And the revenue model is we’re SaaS based cloud product, so we sell annual license subscriptions to all our customers, as well as providing additional integration services on top of that to help our customers get their data into Zegami.
What challenges have you faced on this journey so far and how have you overcome these?
RN: I’d like to say the biggest challenge we’ve faced in building Zegami is really identifying that product market fit. We spun out from the University as a cool piece of tech. But it was really struggling to find a bit of a home. Who are those people and what are those use cases where you get the most value from Zegami. We are not like other competitors or that you could easily replace Zegami with another tool. It’s really been a process of working really closely with our customers and really having a good close look at the data that they are generating and pulling those two together to really help us to understand, who has the greatest need and who can get the biggest value from a tool like Zegami. I don’t know if you ever get to that destination, but we’re closer than ever we’ve ever been before.
Can you tell us how Covid-19 is impacting the business and what steps you are taking in response?
ST: It is a devastating disease and has a lot of impact on patients and also on the businesses. But for us as a computational business, it’s fine, because we can run really efficiently. We work from home, so it’s really full steam ahead for us in terms of development and because we’re fitting into the category of things we can help in. We’re really trying to work with people in the Oxford ecosystem and beyond to help develop systems that help understand the disease.
So, for example, we have a project where we are looking at x-rays to diagnose, whether they are Covid-19 or healthy or pneumonia, and we’re working with various groups down to get more data to try and use that as a prognosis tool. And there are also other projects we are doing that has a potential for Zegami, more at the cellular level and to understand the disease.
Can you tell us what it means to you to be an Oxford University spin-out?
ST: So being an Oxford spinout is extremely exciting. I have learned a lot about business and how it works. It’s in some respect a bit like academia, where you do experiments and some things work, and some things don’t. We’ve got a great Oxford ecosystem. It’s really good to be able to collaborate with different groups around the University on the product. It’s really great to see your ideas rolled out as a fully professional, well-rounded product- that’s a real thrill. I think it’s really great to see all your hard work put into place, with users using it and getting real insight from their data, and that makes it all worthwhile.
What’s next for Zegami? What’s your vision for the future?
RN: In the short term, it is to really get Zegami into the hands of clinicians and end users, so they can help make more informed, embedded decisions based on AI, but also allow them to better explain the output of AI. That also has a secondary use within the regulatory side of things, so that regulators, when they are coming to audit any kind of machine learning system that has been developed, can have full visibility over the data that has gone into making up that model.
But I think, ultimately, where we’re really heading is back to our roots, and what we originally started out to do, which was to just make data more accessible to anyone regardless of their background. And we really feel that tools like Zegami get that sort of technical part out of the way and give them direct access to their data so that they can identify and understand datasets in a way that is not technical and not limiting them to having to learn how to write complex queries or code or anything like that. So it’s about democratising data and data accessibility.
What further investment do you need to achieve this vision?
RN: We are currently actively looking for further investment to really grow the team. We have raised a series of investments over the past few years to really build the product to where it is today. But the next phase for us is on the commercialisation side of things. That’s really what this next round that we’re looking to raise is going to go towards- to build our sales and marketing teams to really allow us to get Zegami out into the world and grow our userbase.