17 August 2026

18 months, $1m, and one goal: turning the UK's Software Security Code of Practice into something businesses can actually use
Today we get to share some news we've been sitting on for a while. Dragonfly, in partnership with UCL School of Management, has been awarded £704,863 (approximately $1 million) in funding from Innovate UK, part of UK Research and Innovation (UKRI), through the Secure Software for Resilient Growth competition. The 18-month project starts this month, and it takes aim at one of the hardest problems in technology decision making: how do you know whether the software running your business can actually be trusted, together, as a whole?
When we started Dragonfly in December 2024, the bet was simple. Choosing software is broken, and the businesses living with the consequences of those decisions deserve independent, honest answers they can act on. So we built a knowledge graph of more than 300,000 software vendors and launched publicly last October, backed by a £2.6M pre-seed led by Episode 1 alongside Portfolio Ventures and Dreamcraft.
Since then we've spoken with hundreds of businesses about their stacks, and almost every serious conversation lands on the same question of trust. That question goes far beyond any single vendor. Leaders want to know whether the tools they've chosen can safely and effectively work together, especially now that AI systems and autonomous agents are being wired into stacks faster than anyone can evaluate them.
The uncomfortable truth is that existing security tools can't answer that question. They assess vendors one at a time, but real-world risk lives in the connections between tools. The data flows, the integrations, the shadow IT, the over-privileged connectors that quietly link everything together. A business can select ten individually "secure" products and still end up with a dangerously insecure whole.
That's not a theoretical problem. Cyber attacks cost the UK economy an estimated £14.7 billion a year, roughly 0.5% of GDP, and fragmented, poorly secured software infrastructure is among the weaknesses attackers exploit most.
We're entering a world where every business will have both human and AI colleagues working together. The technology decisions organisations make over the next decade will fundamentally shape their ability to innovate, compete and adopt AI safely.
Sean King, cofounder and project lead, Dragonfly
In January 2025, the UK Government published the Software Security Code of Practice (SSCoP): 14 principles that set a baseline for software security across supply chains. It's a genuinely useful piece of work. It's also, for most businesses, a document rather than a tool. Very few organisations have the expertise to check whether their stack meets the standard, or to make security-informed decisions when choosing what to buy next.
That gap, between a published standard and a business's ability to actually apply it, is exactly the kind of problem Dragonfly exists to close.
As businesses increasingly rely on complex ecosystems of software and AI systems, there is a growing need for trustworthy and explainable approaches to technology decision making.
Professor Onesun Steve Yoo, Principal Investigator, UCL School of Management
Over the next 18 months, we'll build a secure stack advisory layer on top of Dragonfly's knowledge graph. Concretely, that means:
The result is the world's first machine-queryable trust layer for software. Ask it whether your stack meets the SSCoP baseline and you get an answer with evidence behind it, in seconds. And because it's machine-queryable, it's not only humans who can ask. As businesses hand more decisions to AI agents, those agents need a trustworthy source to check against before they act. We're building the layer they'll check.
Evaluating a technology stack is a brutally hard combinatorial problem. Hundreds of tools, each with its own security posture, assessed against interoperability, regulatory and security constraints that all interact with each other. This is decision-making under uncertainty at serious scale, and it's precisely the territory of our partners at UCL School of Management.
The research will be led at UCL by Professor Onesun Steve Yoo as Principal Investigator, alongside co-investigators Dr Yi-Chun Akchen, Dr Yuri Resende Fonseca and Dr Aras Selvi. Between them, the team brings deep expertise in optimisation, machine learning and decision-making under uncertainty. Pairing that rigour with the real-world data in our knowledge graph, and the real-world questions our customers ask every day, is what makes this project possible at all.
This collaboration brings together cutting-edge academic research with real-world industry challenges to develop practical solutions that can help organisations navigate complexity with greater confidence. We're excited by the opportunity to contribute both rigorous research and meaningful impact through this partnership.
Professor Onesun Steve Yoo, Principal Investigator, UCL School of Management
Our vision has always been to build the world's most trusted intelligence layer for decisions about the tools a business runs on, and this project moves trust from something we describe to something you can query.
The support from Innovate UK is a serious validation of both the problem and the opportunity, and it means this research happens here in the UK, contributing to the UK's leadership position in secure software and AI innovation.
A huge thank you to the team at Innovate UK, and to Steve, Yi-Chun, Yuri and Aras at UCL. We can't wait to get started. Well, technically, we already have.
You can read the full press release here.
Dragonfly News
Catch the latest trends and insights straight in your inbox.
A short call shows you how all of this looks with your tech stack.
Book a demo