Let’s set the scene - my startup is a startup of one. I’m the developer, the designer, the product owner, the ops team and the person who makes the coffee.
If you’ve followed along with the ReplicantGuard posts or the Everblossom intro, you’ll know ReplicantCore and Everblossom have been my main focus for most of this year. I started building it back in February and went all in on it full time.
Over the last few months I quietly put together a funding application to Innovate UK’s Next Wave: Breakthrough Wave 1 competition, which funds innovation across the creative industries. It’s open to any UK business, which means a solo founder like me goes into the same pot as established companies with full teams, track records and - crucially - people whose actual job it is to write funding bids.
The application wasn’t successful.
You’d think that would sting, and maybe it should have. But honestly, it didn’t. I knew exactly what I was up against when I hit submit, and what I got back was genuinely valuable - detailed, independent feedback from people who assess these things for a living, and a much clearer idea of what to do next.
So, in the spirit of building in public, here’s how it went.
The application centred around ReplicantCore, my creative AI platform, and Everblossom, the children’s storytelling app built on top of it - bringing together storytelling, age-appropriate content generation, localisation, character memory and narration into a commercially-ready platform.
To give you an idea of the competition: 700 applications were submitted, 596 went through to assessment (with me being one of them), and 41 projects were funded - roughly 6% of everything submitted. Funding was split across four “Frontier Industries” - Advertising and Marketing, Film and TV, Music, performing and visual arts, and Videogames - and every funded project scored somewhere between 84.7% and 94.3%.
It’s a tough gig.
I didn’t make the cut this time, but for a one-person company in a field of 700, I’m happy with it. The technology, innovation and project plan scored well, as did value for money.
Where I dropped points was on the business side - market and route to market - and, no surprise, on team size.
The feedback comes from independent assessors, and Innovate UK can’t discuss it or add anything to it, so what’s on the page is all you get. There’s a lot in there I’m genuinely pleased with.
One assessor described it as a:
“technically developed platform with a clear architecture and some interesting approaches to safety, localisation and narrative consistency.”
and said the project “shows capability and ambition”. Another said “the idea in this application may well be a strong one”. And one summed it up in two words:
“Highly fundable.”
As a lone developer who’s spent the last eight months building this on my own, that’s lovely to read. When there’s no team around you to sanity-check your ideas, independent experts telling you it’s well built and worth pursuing counts for a lot.
Like a lot of technical founders, I spent most of the application showing the technology worked - and to be fair, that part went down well. What the assessors wanted next was the business half of the story: how it’s different from what’s already out there, what the plan and numbers look like, and who’s delivering it alongside the founder.
The first one is the most interesting to me. One assessor felt the advance over existing AI safety, orchestration and localisation approaches was “only partly demonstrated”. The tech is there - ReplicantGuard, for example, is fully deterministic and explainable with no external models, which is a very different beast from the usual off-the-shelf content filters - but I didn’t spell out that comparison. I explained how it works, rather than why that’s better. Lesson learned.
Another comment stood out too:
The application also provides surprisingly little detail on “Everblossom”, despite it being central to validating the platform’s user appeal and commercial potential.
Fair point. ReplicantCore is the engine, but Everblossom is what people actually see and use. Depending on the audience, some people are more interested in the platform tech and some in the app itself, and I’ve always struggled with that balance. In this application it ended up as a middle ground - when the product should have been the star of the show.
Team size was my lowest score, and I saw that one coming a mile off. The company that submitted the application consists of… me.
From a funder’s point of view, that’s an obvious risk. The question isn’t “can he build it?” - it’s “what happens if he can’t?” When the companies you’re up against can point to a team of ten, it’s hard to compete on that front, and I completely understand why it was flagged.
Helpfully, the feedback also pointed at how to address it. The assessor who called it “highly fundable” said the main risk was that subcontractors hadn’t been identified. Being a solo founder doesn’t mean doing every single thing yourself - it means being clear about who you’ll bring in, and when.
Another assessor noted there wasn’t much detail on the person applying. I did mention myself and my history, but I didn’t go far enough. I’ve been building software for a long time, for companies ranging from big tech to small startups, and that’s exactly the kind of track record that reassures someone reading about a one-person company - but I need to push that more.
The good news is that almost all of the feedback is actionable - and most of it is about how the application was written, rather than what’s been built.
There’s another side to this that I want to be honest about.
When I started building this back in February, I wasn’t on a client contract, so I went all in. Full days, full focus, and a lot got built. Then, around three months ago, a full-time client project came along. Bills are bills, so I took it, and naturally that’s where most of my attention has gone since.
Going from founder mode to side-project mode is a much bigger shift than I expected. When you’re full time on a startup, you can hold the whole thing in your head. When it’s squeezed into evenings and weekends after a full day of someone else’s problems, a lot of that time goes on just remembering where you were. Progress slows, and slow progress has a way of making you question the whole thing. I’ll admit I’ve had a few “should I just call it a day?” moments.
What stopped me was realising that nothing about the idea had changed. I still love it, and I’m still excited by the tech. The only thing that had changed was how much time I had for it - and that’s a scheduling problem, not a reason to quit. Full-time weeks are for building the platform; side-project weeks are for smaller, sharper things. That’s something I’m going to plan for better in the future, and for other projects too.
The most useful thing to come out of this wasn’t the score - it was the roadmap. As a solo founder you don’t often get experienced people to give your business a proper once-over and tell you, in writing, exactly where the gaps are.
So the focus is pretty simple - get Everblossom into more hands, line up the partners and supporters, and keep building out ReplicantCore alongside it, with a clearer idea of what to do, and what not to, when bill-paying client work comes round.
Nobody doubted I could build it - they were asking to see the business behind it. Next time I apply, I intend to show them. I’ll be sharing how it goes here.