Sundog Bio News | June 2025
Good morning,
We’ve spent June talking to potential users. Principal investigators, postdocs, bioinformaticians and industry scientists have all shared their feedback and challenges. They’re helping us to develop a truly compelling vision for the product; something that will change the way biologists work, for the better, forever.
Excitingly, every single research scientist we’ve spoken to wants to stay involved. They’ve seen how Sundog can help them with the problems they have today. We have both a Scientific Advisory Board with whom we can develop the product, and a range of real-world use cases we can build for.
Alex Mitchell | CEO Update
I want to share the big themes - the most important things we learned. And I want to share with you the voices of our potential users and customers, as they’ve thought with and given feedback on the software we’ve built. This is the work I love. So happy to share it with you today.
Collaboration is (even) more important than we thought
We decided to focus on collaboration early in our product development. We knew that biologists work closely within their labs, and that the analysis phase of an experiment was usually the most isolated. Our assumption was that scientists in the same lab would want to collaborate with their small, trusted team. And that’s important.
So this is like, image Google Docs. — Principal Investigator, University of Bath
Almost immediately, our participants started talking about the more complicated cases where Sundog would change their remote collaborations:
We've got a co-supervised PhD student starting, so we could look at data together on a call. I like the idea of annotating. […] No need to put a hard drive on a train to London. — Principal Investigator, University of Bath
Those collaborations extend not only across labs, but across disciplines. Biologists collaborate with pathologists, bioinformaticians and clinicians, and there’s no easy way for these geographically distributed groups to share expertise:
Two people being able to use this at the same time, or multiple people, is fantastic. We've needed to consult with pathologists because we're not pathologists, and it's all done on a shared PowerPoint presentation. So this is fantastic. — Principal Investigator, Imperial
Biologists want better in-person collaboration
The case for asynchronous collaboration is clear. Biologists work across geographical sites, across time zones, and everyone (always!) has a packed calendar. Something that surprised us as we spoke to our potential users and customers, though, was their desire to use Sundog to support their in-person conversations; in their lab meetings, for instance, or in training new scientists.
When I have meetings with the lab, often they'll have a graph and I'll say, ‘oh that's interesting - what do the cells look like?’ Then we'd have to go and walk over to the microscope [...] but something like this, you just click on a link and you can just start exploring together. — Principal Investigator, Imperial
We talk about wanting to be the ‘Canva of life sciences’; a simple tool that lets anyone produce work they’re proud of. Our conversations in June open up the possibility of being the ‘Miro of life sciences’; an online tool that supercharges in-person collaboration.
“How this works currently, it[’s] over Zoom calls, and the images are horrendous. You're letting Zoom pixelate everything really really badly and you're losing resolution” — Bioinformatician, EMBL-EBI
Scientists appreciate simplicity
We’re betting on simplicity at Sundog. These demo sessions were important in helping us to understand whether that was right for our target market. Both Tom and I went in with a degree of trepidation that we’d be asked to implement ImageJ in the browser.
Omero does too many things. — Freelance pathologist
Fortunately, no. We don’t have to be the only tool a scientist uses to analyze their data. What we want - and what biologists want - is for sharing images, understanding them, and communicating results to be a joy. We can’t wait to start building this better world.
We could have done with this for some of our figures, which have been painful to curate. — Principal Investigator, MRC Laboratory of Medical Sciences
Tom Armitage | CTO Update
It’s been great seeing the response to the demo. It’s also been great seeing how the bets we made - on what to demo, and how much work to put into it - have paid off.
Demoing the product with a representative sample file (our 3.8gb tiff) paid off. Proving we could work with realistic files, in formats biologists understood, meant we built trust early on, and kicked off interesting conversations about the file formats our scientists work with. For some, our sample file was similar to the shape of imaging they do (a large single tissue with many layers). For others, it was very different - but we could talk about those differences with them in confidence. (In terms of size and complexity… it was very much in the middle of the requirements our test users told us.)
Building the end-to-end story for real gave interviewees the best chance to understand the whole product. We showed a whole end-to-end journey through the application, stopping at the main ‘bullet points’ in the narrative for feedback. Because each stage of use builds on the previous, we started seeing our users ‘thinking with the product’. Suddenly, a lightbulb would flick on in their head, and they started reasoning about how they would use this in their labs. And they did this in a way that was completely in line with what we hoped the vision for the product would be: it was exciting to see how entirely they had come to understand the vision. Every feature we built for real - whether it was file ingest, realtime multiplayer collaboration, or output rendering - helped them see their own use cases, and share better, more detailed feedback with us.
Trusting ourselves to show a minimal demo was worth the pain! The early, minimal state of the product turned out to encourage engagement from our interviewees. They would spot quick ‘wins’ to suggest, and enthused by our response, move on to feedback on more complex issues and pain points. The software is still too minimal to be a viable product, but it’s a highly viable demo.
I was concerned that our web UI would be too simplistic for our scientific audience. In fact, our interviewees responded enthusiastically to our interaction design approach. Sundog never came over as a “Fisher-Price” product. Rather, multiple interviewees were enthusiastic about the ease of use, especially when they considered sharing images with less technical colleagues, and we got an unprompted I love how easy it is to use! out of one interviewee. So we’re going to keep investing in straightforward, high-quality, consumer-grade interaction design.
I also started exploring what you could call our “AI/ML vision” with our interviewees, seeing how they worked, and how they responded to some questions.
A few people - funders, peers - have asked us what we’re doing about “AI” in Sundog, and I’ve been deflecting the question whilst I work out how to phrase the answer in a way I’m happy with.
The answer is straightforward: we’d like to work with the Machine Learning techniques our scientists are already working with, and have been for many years.
ML tooling is used for tasks like image analysis and cell segmentation. Many labs have pipelines they’ve established over years and already are happy with. Why would we try to compete with them, or re-implement them?
Sundog’s focus is on scientific collaboration - as Alex writes above, our interviews have only made us more certain of that. And some of that collaboration is with machines. Given that, it should be as easy for a bioinformatician to use a Sundog API to get an image from Sundog, run it through a local tool, and push an annotation layer into the image, as it is for a scientist to make annotations by hand.
To put it another way, my AI strategy is from sundog import client. We supply an API and tooling, and let users integrate with existing tools. And one of those users is us - this is a great mechanism for prototyping further ML functionality; perhaps there will be tasks so ubiquitous it’ll end being worthwhile building them in ourselves.
We’ve been talking a lot about our demo, and we’d love to share it with you as well. Just reply to this email, or email alexandra@sundog.bio, and we’ll get time in!
In July, we’re going to step away from product development - building - and return to sketching, thinking, planning. We’re going to take everything we’ve learned, and use it to recalibrate our product vision, strategy, and roadmap. The demo promised the value scientists would get from using our tool, and we want to be sure that from the very start, Sundog is delivering that value.
We’ll share our vision with our Scientific Advisory Board - and start commercial conversations with our earliest customers and collaborators.