We did not set out to build a company that predicts how a piece of creative will actually land with consumers before it ever goes live.
Ten years ago, that sentence would have meant nothing to any of us. Today, I'm co-founder and CEO of Dragonfly AI, and it's a question the team and I spend most of our working lives trying to answer.
My route here has never been a straight line. I studied cognitive science – psychology, computer science, mathematics and philosophy, all pointed at one question: how does thinking actually work? Then I DJed for several years, borrowed money from my parents for a music studio that never became the empire I'd hoped for, and eventually taught myself to code.
Looking back, there's still a thread: a fascination with how humans think, paired with a genuine love of technology. Everything Dragonfly AI does today sits in that same space – human behaviour, data and technology, brought together to show brands things they didn't know they didn't know.
Before Dragonfly AI, I co-founded Black Swan Data, a business built on predicting what people would talk about next from what they were already saying online. It taught me that the idea I actually wanted to spend my career on was prediction, not analysis after the fact. Dragonfly AI is where that idea properly took shape.
I first came across the technology that would become Dragonfly AI at a university show-and-tell, listening to professors explain work far smarter than anything I could do. Professor Peter McOwan and Dr Hamit Soyel, at Queen Mary University of London, had spent years building a computational model of how the human eye decides what to look at first, developed originally for robotics.
It didn't need an eye-tracking lab to predict where attention would land – it simulated the biology directly, and still does, validated at 91% relative accuracy against professional eye-tracking panels. I knew how much revenue CPG and FMCG brands lose to pack designs that never quite land the message. We had the tool. We didn't yet have the business – so we built Dragonfly AI around it.
We didn't get it right immediately – building a company is always more like climbing a mountain than following a map. What changed things was scale: putting the technology against thousands of pack designs rather than a handful made the patterns impossible to ignore.
Today, Dragonfly AI's platform is used by more than 50 global CPG businesses to see their creative the way a shopper actually does – a validated, patented piece of science turned into infrastructure a brand can rely on.
The belief underneath Dragonfly AI is a simple one: human response can be predicted before it happens, not just measured after the fact once the budget is already spent. That's not a new idea for me – it's what first pulled me into this industry – but Dragonfly AI is where we've been able to apply it properly, to creative effectiveness: not just where a person's eye lands on a pack, an advert or a product page, but what they remember and how they feel about it afterwards.
That question matters more now than when I started. Gartner forecast that 30% of outbound marketing messages from large organisations would be AI-generated by 2025, up from under 2% in 2022 – a milestone we're already past. Brands are producing more creative, across more channels, faster than any team can realistically review.
More creative was meant to mean more chances to get it right; mostly it's just meant more going out unchecked. Attention is still the first filter any piece of creative has to pass, but it's only the start – what people remember and how they feel about it matters just as much. Dragonfly AI is one of the few businesses built to measure all of that before a single asset goes live.
I still don't have a tidy answer for how a cognitive science degree, several years behind DJ decks and a failed music studio led here. What I do know is that Dragonfly AI has come from the same place every business I've built has: a moment of ‘wait, why isn't anyone doing this?’, followed by a lot of hard work to find out if that instinct was right. I've been lucky enough to spend the last few years asking that question about creative effectiveness, at a company built to answer it properly.