Defence Graph

A map of the UK defence innovation landscape you can ask questions of, rather than only look at.

What this is

inink built an interactive map of more than two hundred organisations, programmes and frameworks across UK defence innovation procurement. It is a good picture of a complicated landscape. A picture is where most founders stop, because the next question is always specific to them.

This demonstrator loads that landscape, together with the knowledge in MilUX's Foothold pack and a slice of the Tally directory of UK defence companies, into TrustGraph. It then puts one fictional founder inside it, so you can watch someone actually interrogate the thing.

Three real offers, one fictional customer

inink

The ecosystem map, and SC-cleared BD, sales and MarComms support. The founder in this demo got their positioning and BD narrative here.

MilUX Foothold

The free vault pack a defence founder runs their business on. The landscape knowledge in this graph comes from it. The paid diagnostic sits above it.

TrustGraph

The open-source graph and retrieval platform underneath. Apache 2.0. It is what turns the two knowledge sets into something answerable.

Four questions to ask it

  1. Competitor analysis Who else is going after this? Try: Which companies in the South of England compete with us on uncrewed air systems, and which are really a route to market rather than a competitor?
  2. Go-to-market planning Which route, which framework, which door? Try: What are our realistic routes to a first defence contract, and which needs a framework place we do not have?
  3. Stakeholder mapping Who do I need to engage, and in what order? Try: Which organisations should we approach, in what sequence, and why that sequence?
  4. Customer discovery What is funded, who runs it, who is competing? Try: Which funded programmes are relevant to tactical ISR and autonomy, and who owns them?

An honest note on how the graph is built. The landscape spine of this graph is built deterministically: every relationship between two real entities traces back to a stated field or an explicit link in the source, and nothing is inferred. Relationships involving the fictional company are authored. Where a language model is used to extract further detail, it is stochastic and it will get some things wrong, so treat any surprising relationship as something to check rather than something to rely on. Provenance is shown so a wrong edge is traceable rather than invisible.

Built by MilUX with inink and TrustGraph. The founder company, its staff and its engagements are fictional and any resemblance to a real company is unintended.