Discussing the future of deeptech with AutoFill CEO Gideon Richheimer
In our latest exclusive interview, we chatted with AutoFill Technologies CEO Gideon Richheimer about the future of deeptech.
What is your background and what were you doing before you founded Autofill Technologies?
Before AutoFill, I co-founded several other startups, including a software product company which became a market leader for data monitoring and was later acquired by a spin-off from Mitel. I even ventured into a B2C startup, but quickly realised that complex and research-driven products focusing on the enterprise and B2B are more my cup of tea.
Generally, my role at these companies was behind the scenes, making sure products were relevant for customers, rather than merely nice-to-have solutions. I also made sure that we focused on the bigger picture, planning and strategising for the future ahead, in order to achieve our business goals.
After I exited my last company, AutoFill as an idea began to take shape in my head, following visits to some of my old partners and customers’ sites, where I saw how inefficient they were at inspecting their transportation fleets.
What difficulties are there when raising funding and how much does it help to have Rick Belluzzo as an investor?
We received relatively quick interest from investors at the beginning of our journey. We didn’t have our backs against the wall in the sense that we didn’t have to raise money immediately because the cofounders, including myself, Stefan, Daan and Luc, were able to fund AutoFill ourselves in the early days.
So, we were a bit more selective in terms of the path we wanted to take. For instance, some of the early investors, although I can only speak highly of them, didn’t come from the engineering field. We believed that, for our initial external fundraise, the organisations would have to have a deep understanding of the world we’re active in.
So, we kickstarted in reverse order when we were selected by Deeptech Labs to be part of their accelerator programme. This is an outstanding VC fund and accelerator, backed by Arm and the University of Cambridge. Twice a year, DTL invests £350k in a small number of deep tech post-seed startups. These companies are then embedded in a powerful network of successful entrepreneurs, expert practitioners, leading researchers, and deep tech organisations worldwide. AutoFill was included in the first cohort, and we believed this was the right way to go due to DTL’s expertise and knowledge.
After it was confirmed that we were part of the Deeptech Labs programme, conversations with lead investors that had already begun prior to our engagement with DTL deepened.
We’re proud to say that one of our investors and advisory council members is Rick Belluzzo, former COO of Microsoft. Rick was interested in AutoFill because he sees that there is an opportunity to take the technology to multiple industries.
Having Rick on board is amazing for us. He is very engaged in what we’re doing, and he has been a huge asset for us from an advisory perspective.
What are the current difficulties for inspection workflow processes for the automotive and rail industries?
Currently, the vast majority of object inspections in the automotive and railway sectors are performed by humans. This clearly comes with huge risks, both to safety and financially. Existing automated solutions, meanwhile, are costly and require substantial space and infrastructure to function (often a facility the size of a car wash).
How can Artificial Intelligence help overcome these barriers?
Through AI, you’re not limited by time. Computers don’t clock in and out. Inspections can be performed at any hour of the day, fitting perfectly to any operational process.
Additionally, AI adds objectivity to the inspections. It is an objective analysis of what the situation is. Therefore, in our opinion, we consider it to be a true inspection. It doesn’t take into account weather or light, but also not mood or emotion. For instance, if you’re a doctor and you need to make an inspection of the skin of a human being, emotion is also a part of the assessment, where you also have to analyse how someone feels.
But with analysing the state of a vehicle, you don’t need any of that. It is completely evidence-based. The objectivity of a report is something that is extremely valuable because then you can create operational consistency.
What trends are you currently seeing in the deep tech sector?
I wouldn’t say deep tech is a buzzword, but it’s pretty much out there now. But I think a lot of what we’re seeing is similar to what AutoFill’s doing – bringing research to life.
True deep tech is, many times, born out of academic research. I’m also observing that a lot of non-deep tech companies (e.g. investors or market organisations) have a good understanding that deep tech is not only valuable, but it requires time. It’s not something you can just press on and it’s there. It requires training of the system and situation, which is a whole other topic on its own.
When deep tech ties into hardware, that’s the ideal setting. When you have trained data sets piping through a piece of hardware, that’s where you start seeing an enormous change of existing operational processes. This will be a big focus for companies and industries in 2022.
The calculation power of these types of software sets can sometimes be a bit intimidating because it makes very smart analysis very quickly. There’s a scary side to it, but also a good one because it means you automate certain jobs that have huge safety hazards, like military surveillance in very difficult terrains and areas, where you need to take into account potential human casualties. Those are really good adaptations. It is amazing to see the power of what the technology can do.
What does the future for deep tech and Autofill Technologies look like?
One of the main topics that we focus on is extending our embracing of additional, exciting technologies which we can fuse together to achieve higher accuracy and evidence-based objectivity. Our goal is to reach a certain data threshold. Once we’ve had a certain number of images processed, then the technology becomes intuitively smart. That’s when you can identify really interesting patterns.
By having better pattern recognition, you can start following the roots of issues or events and not just react to a situation that may have occurred but go way back in a chain of events and act proactively.
Take railways, for instance. The way rail infrastructure is currently inspected is simply by looking at it. But once you start to understand how the inside of the steel is working and how it’s continuously, dynamically changing, we see a future where you don’t even need to look at the outside anymore. You’ll be able to look at the inside based on mathematical calculations. You can really be ahead of the curve and therefore create a much safer environment. This is one of our main goals.