Meet Tyler Edwards: The former MI5 and GCHQ agent who just raised £2m for his AI systems!

0:00Tyler Edwards:Hello and welcome back to Scaling Europe show presented by Deal. I'm Sid Johnson. Today I'm joined by none other than Tyler Edwards, co-founder and CEO of Overminds, who's got very exciting news today. What's going on? What's the news? Today we are officially just coming out of semi-stealth. We're announcing release of our product and the seed round that we've just closed. So yeah, it's a big day for us. Amazing. Tell me about the seed round, how much, who's getting involved? yep so we raised two million in total um and it's led by osney capital so they are a cyber focused fund based out here in the uk um supported by a number of other funds so portfolio ventures uh 14 peaks uh endurance ventures and antler can't forget antler um a range of different funds uh i guess the objective here was to stack the team quite well so osney very cyber focused we're obviously building in the security industry and we also wanted some generalists to come in we are an early stage team we're first-time founders we wanted to surround ourselves with some yeah really punchy expertise as we go on this journey nice well look for those who aren't familiar with overmind as you said you've kind of been operating in seven self what's the 32nd pitch what is it that you guys have built yeah yeah so i'll set the scene a little bit so you're not really going out on a limb to say that agentic ai is clearly on the rise at the moment and obviously most teams are using in quite restricted capacity often like low risk use cases and this is largely because we lack the necessary tools and infrastructure to ship these things at scale at the moment.

1:25So really simply put, Overmind is a runtime security tool that watches your agent, understands its behaviors and patterns of life, captures its mistakes, then uses fine tuning to refine and specialize your model.

1:37Tyler Edwards:Amazing. And you mentioned that you're all first time founders, but this is like a really interesting team that you've got. um i think you know your background is kind of building ai tools for mi5 mi6 uh akat is like doing some machine learning stuff at places like monzo enlist your cro sam has been working in go to market three different unicorns this is like a cracked team right can you touch a bit about let's kick kick off with your experience first right because you've got the really interesting background how did you get into this space why have you decided to build this solve this problem yeah Yeah, of course.

2:13So I was in the defense space for about seven or so years, came in initially as an engineer, building sort of offensive cyber capabilities, worked across like the Russian, Iranian, Chinese, all these sorts of bits. Towards the end of my time there, they started trying to see how they could deploy what was machine learning at the time, but eventually AI into sort of intelligence analytics and other defense capabilities. Famously, government is not great with tech uplift, and they hit all sorts of blockers. They really, really struggled with scaling and implementing it. But weirdly, the exact same problems existed in the commercial world.

2:49This sort of massive ambition to deploy this really powerful tech that is utterly transformative, but people really failing slash struggling to do so. So actually, some of the tech that we used in the defense space, we are deploying into Overmind as a little bit of our secret sauce. to get it all working.

3:08Tyler Edwards:Really cool. And how did you know your two co-founders? Yep. So we actually met through the Antler Accelerator Scheme. So myself and Akat, we met day one. Annoyingly for Akat, the day he landed on the scheme, I immediately grabbed him. I was like, I've seen your background. I think you're amazing. Let's work together. We stayed married throughout the entire thing. Towards the end, we realized that we were two massive nerds that could build incredible stuff, had never sold a thing in our lives so didn't really make sense for us to run like a competitive go-to-market motion and thus sam came in so one of the partners antler introduced us to sam he was just absolutely incredible um he quit his job before we had like funding before like any real sort of traction um but it was what sort of helped build confidence in him that you know he understands what we're doing he sees the vision and he's ready to go amazing and uh how was that process how was it going through antler what was all that like yeah it was i i always say i think this has been more stressful than than being an mi5 um different kinds of stress but yeah far more stressful it's really cool uh met some brilliant people the partners there have been amazing at supporting us um the the people that i met and worked with there are still my friends now um and i obviously come from an environment where i didn't exist where i didn't have social media i didn't have any sort of public presence um so it really allowed me just to like come out of the secretive intelligence world into the arena that is yeah startup founding amazing and let's talk about the round you know two million dollar seed round what what unlocked it you know what gave the investors that you have got now on board the conviction that you are the team building the product that they want to back yeah i think it was a compliments of things like you said the the team itself is is impressive um you don't really see many teams in this space with with our makeup But usually they're highly academic, not much in terms of sales and enterprise.

4:59And you certainly don't see many teams coming out of the defense world and founding. I think obviously most famously Darktrace, obviously founded by previous members of GCHQ. I think for them, the unlock was the way that we looked at the problem. I think AI security at the moment is quite a nebulous area. It can mean a number of different things. We didn't want to build like another enterprise SaaS tool. We just didn't think that was the way to crack the problem. we wanted to build something for developers for engineers the people that are being ambitious and brave and building really cool products um so i think that's what helps sort of draw some of that out and of course osney being cyber focused i think they enjoy that we're trying to pull away from the pack and not just sell another like monolithic enterprise ai security tool really

5:41Tyler Edwards:cool and i guess like what is the pitch to companies you know when you when you're going to prospects and clients what are the real pain points that you're saying look this is what we're going to do this is what we're going to solve or protect and what is it that they're finding convincing yeah well i think at the moment it depends on the persona so for security focused teams it's a lot about that ambition piece so you want to ship all of these really incredible products but you know they're really difficult to manage to observe to govern behaviors so shipping a tool that runs at runtime it understands how your agent works it understands the pattern of life gives you that really hands-on intrinsic understanding of the product you're building for i guess less security focused teams it's kind of one of competitive advantage i guess we've come to the point now where ai is a commodity saying you use ai is like saying you use the internet like it's it's it's a given so really what pulls teams apart is how they execute how do they use ai um one of the things that our product unlocks which i think is is really cool is it enables teams to ship specialized models so uh for context if you are doing kyc in a bank for example your model doesn't need to know the full works of Shakespeare.

6:45It just needs to know how to do what's job incredibly well. So we can help teams ship these really fast, specialized models that can out-compete generalist models. So it's not just a solve for security. This is about how do you unlock the next era of authentic AI.

6:59Tyler Edwards:Super interesting. And can you touch on, I guess, any of the success that you've seen so far? Have you been, you know, you've been in semi-stealth. What does that look like? Have you been running it with clients? What success have you seen? Have people really seen value from it already? Yeah, so during that semi-stealth period, we kind of worked, I guess, like a pseudo consultancy. So we would find teams doing really cool and ambitious things with AI, and we would kind of work alongside them to understand what the problems were and how we could help. So the number of teams that we've helped ship products now, so one team, they were doing, I guess I need to be slightly aloof about it, they were shipping a product that extracted data from customer input, and then an agent was cascaded, some form of automated workflow.

7:38So at the time when we started working with them, the agent was incorrect 54 % of the time. So it wasn't a particularly reliable product. After we came in, we got those numbers up to 94 % just through these sort of fine-tuning observability pieces. So really, it was a bit of a symbiotic relationship. We wanted to prove our technical hypothesis that it actually worked, but also make sure we were building something that had a repeatable problem statement, it could scale, and was useful to the end user.

8:06Tyler Edwards:Amazing. And what's next? you know you've raised this money you know it's the start of the year you're going to deploy the capital ramp up scale what does that look like yeah so i guess the next step is kind of more more of the same we still have a lot of technical hypothesis we need to prove out this is still a very very nascent market we've just really touched the surface of what's happening in ai you know things like claude bot that came out recently constantly new models being released the ground is moving underneath us so the first thing we want to do is build a very strong technical team that is excited by the industry and is keeping tabs on what's going on and is sort of able to react to these massive changes we have our us trip planned and coming up we want to explore the market there a little bit obviously very different to the market that exists in europe at the moment and then hopefully we want to start expanding beyond just the sort of bottom-up dev rail customers at the moment and start exploring more enterprise adoption very different sales motion from from selling to engineers but that's where some of the really exciting projects are happening yeah i bet and that's that's how you crack the big contracts the higher attention the real uh money holders i guess when you're going out to the us what's that going to look like is that just trying to find people you can sell to is that understanding whether there's appetite for market expansion are you going to be trying to meeting any investors what's that going to be like yeah it's going to be a mixture of the two so we don't want to sort of go all in and decide that we're going to wholesale ship across to the us but we do want to get an understanding of what the landscape is like out there i think for the area we're building in so dev tools ai infrastructure is a little bit more crowded than in europe um the sort of things are spinning up and getting hyped and coming down again like within the week um so we want to get an understanding of yeah what what that looks like in the us take some of those learning backs learnings back to the market in europe as well um from the investor side so yeah we're obviously already thinking about our series a is something we'd like to do maybe like a year and a half a little bit a little bit longer don't quote me on that we'll see we'll see what happens but we want to start exploring those those u.s investor relationships as well kind of explore um how the cell lands over there and what the teams can do to help us scale in that new market um and we just want to start pipping on people's radars you know really we've been going at this for well since july of last year so we've done a lot very quickly but now we need to start building our profiles you know get people knowing who we are what we're what we're doing the new kids on the block so yeah it's all part all part of that and what advantages do you think you have over the classic sort of like silicon valley background which is often i mean either x big tech or x uh researcher you know you've got a very unique background coming from a sort of like defense base what advantages do you think that gives you over the more traditional tech background yeah i do think that so when we were initially sort of pitching for for investment um one of the bits we were told is investors are going to see the same shape of team over and over and over and over again um and what you're going to need more of now is that initial differentiation like what is the story that makes you really really cool so even just from that narrative piece i think coming from the background i do not just coming from the standard sort of cadence of like big tech or the same universities and institutions i think changes massively um i think from a technical perspective in the the the background i come from you're constantly being asked to do impossible things in impossible circumstances either you know build some technology in 24 hours before something catastrophic happens you know these these sorts of things i think having that kind of pressure is very different to building commercial products um there is obviously that that commercial pressure there but it manifests in a very different way i think coming from that environment um as much as i i do maintain this is far more stressful i think i was more prepared than i thought i was going to be um for some of the challenges that that i've come up against i think you know i i've had to pitch to five-star generals of of not so friendly countries pitching to vcs was much easier yeah and i and that's that's super interesting how did you find that that that process because you know coming from a background which i guess is not everybody in big tech everybody who comes from you know one of the leading model providers or you know like your co-founders who've been working in tech for a long time or specifically like tech is in the vc bat world how did you find that transition from i guess was probably a very different environment probably very structured and maybe high stress and pressure but you're working for an incredibly old well-established organization versus i guess your co-founders who are much more used to this sort of like scrappy tech life how did you find that transition yeah i think there there were some things that did remain the same and some things that that that weren't so um in the in the government world commercials all those sorts of things you just don't care about because it's not it's not part of your of your sort of daily pattern um so i think having to build up my commercial acumen sort of how i walk the walk talk the talk in front of customers and vcs was one of the big pieces um in terms about the scrappiness again you You are having to be incredibly scrappy in the intelligence world because most days you're facing problems that no one has faced before.

13:14Like an entirely new situation has arisen with an entirely new set of variables and circumstances and you're just being asked to solve the problem on the spot, which is very much what startup founding is. You all of a sudden have all of these massive problems that you don't really quite know how to solve yet, but they have to be solved. So you just find a way. I think it's that sort of nature that translated very well across to the startup world. I would say the commercial bit was hard. I learned very quickly. I was not a good salesperson, which is fine. You know, these things you can learn. But yeah, Sam has really helped bringing him in to build some of that acumen.

13:51Tyler Edwards:Yeah, it's amazing. Super interesting. As you step into that sort of like that founder role, you're all of a sudden responsible for everything. You're not going to be great at everything. It's important to learn where your skills. kind of are and where they're not yeah and how did you find the pitching you know like going out to investors trying to raise money because that in itself is a sales process to an extent you know you're selling yourself you're selling the business um how did you find that yeah uh weirdly it was fun to start let's say um i think me and sam you know we hadn't known each other very long and all of a sudden we were spending like 90 hours with one another just constantly pitching to investors I think we ended up pitching to I want to say like 100 different funds throughout October.

14:35But it was an interesting process. I think it was sort of quick fire learning. You have to really refine your pitch, really refine your understanding of the landscape, really refine your understanding of the tech and all these sorts of pieces. I think it helped to reflect back what we were building and what we were doing because you would go through these pitches. The investors then kind of mirror them back to you. and sometimes you think oh that's not quite what i wanted to put across in terms of the team the product our ambition um so really think that whole process was quite useful in us really understanding like who we want to be what we want to build um but no it was uh fun maybe not fun but it was it

15:14Tyler Edwards:was it was an interesting process let's say yeah an experience uh looking forward you know you've talked a bit about how you're going to deploy that capital what is it you think you need to achieve right before you you get to the end of the year before you get to that next fundraise are you thinking in terms of revenue milestones customer numbers or is it more about delivering a world class product how are you balancing that yeah i think it's a mixture of the two i think for a team like how it's building infrastructure and dev tools revenue can be a bit of a sort of catfish metric if you only have like a couple of people using your tool but they're paying a lot of money that's not indicative that you've built something transformative we want to build something that developers absolutely love that engineers really really love we want to initially build that that scale of um usership and then we'll start really thinking about about revenue conversion i mean sam again his his whole drive is like you know how do we commercialize this thing how do we drive revenue um so i leave him to to really focus on those bits but i think for me someone that was a career engineer i know that these tools only work when like tens of thousands of engineers love them I mean, teams like, you know, Auth0, SNCC, those sorts of people, that's how they started.

16:24They had massive adoption amongst engineers, developers, all the rest. And then they started converting to these sort of very successful revenue models. And we want to follow a sort of a similar cadence. I think because the market is so busy at the moment, you could just release something and the hype around AI would draw attention and eventually people would churn. That's not the company we want to be. We want to enable the future of AI. And to do that, we need to be considered and careful in how and what we build. So we want to build something that people stick with.

16:57Tyler Edwards:And what's your plan to get those engineers on board? To see that sort of the same go-to-market model that SNCC and Auth0 saw? What's your way into those engineers? Yeah, well, part of this recent launch has been a whole rebrand of the company. So as I said, when we started, we were very much focusing on like enterprise SaaS, like enterprise adoption of AI, these sorts of bits. So the way that we looked and field was not appealing to an engineer. So part of it has been taking the shape of a company that engineers will resonate with, either in sort of our cultural references, you know, our logo and our branding is more akin to a dev tool.

17:37The second part is kind of what I said, just building something that is genuinely useful. it exists and integrates with the places people are building it's not bashful we want to be brave and ambitious with what we want to enable people to do we want to be opinionated in what we think the future is going to look like and technologists, engineers, scientists, researchers respect that they don't want to sit in yesterday's tech stack they want to see what crazy things they can do so a lot of our narrative is like we don't want to get in the way we want to take off the brakes give you cool tools to build stuff and i think trying to resonate and land that kind of narrative is is yeah a good initial play but really simply put build something engineers like makes their job easy and they will stick

18:20Tyler Edwards:with you yeah amazing and it's yeah it's gonna be a great strategy to build product people love and the rest will follow uh well look tyler thank you so much for joining me it's great chatting i'm super excited to see what you get up to this year i think what you're building is it's really interesting and i'm a huge fan of the the building the the founding team that you've together um so best of luck with it thank you very much it was great to meet you