Origin — A twenty-year vision to close the gap

THE IDEA

Every hub begins the same way: one person holding a vision for a group. The engine began that way too — in 2007, inside consumer finance, long before it had a name.

2007 — The principles

The recognition: when one organisation must make decisions for many people on a shared journey, the truth it needs lives with the people — and it surfaces when each person works independently, in private, and shares their position with the centre. The first build ran it at scale: private contributions, gathered to a common hub, composed a shared view both sides could act on — and decisions moved faster because of it.

2014 — The structure

The innovation field showed the same gap at a larger scale. Collaboration ran through established partners, while the breadth of research, entrepreneurship and capability across the wider landscape went unseen. The second build gave the vision its structure: analyse a programme brief, read the published data of a whole landscape, and shortlist the organisations best placed to collaborate. The principles held. The technology of the day could not carry them further, and the third build came twelve years later.

2026 — The capability

AI changed what could be built. Where the second build read what organisations published, the third gives each member a Personal AI to speak with in confidence, in their own words — and composes what they hold, at scale, into the group’s collective position. Twenty years after the principles, the capability arrived: Horizon Intelligence.

THE RECORD

Twenty years of work, in three sectors entered one at a time. The consulting came first; the technology came out of it.

THE CONSULTING

The work began in the rail sector, convening industry workshops at national scale for its innovation programme — the first encounter with the discipline of carrying what a group knows to the many. It went on to the Science and Technology Facilities Council, translating the capabilities of the UK’s ISIS particle accelerator into industrial propositions and designing a collaborative programme for metal additive manufacturing and the aerospace primes, structured so every participant kept their IP. At UCL’s Bartlett, the world’s leading school of architecture, it was research partnerships and collaborative grant applications, from first conversation to submission.

Then the innovation system itself: programme work with Innovate UK’s Transforming Construction and the ICURe accelerator, taking university research teams from laboratory to commercialisation, a strategic review of the programme’s own operating structure, and a feasibility study opening early-stage STEM innovation to impact-led investors. And a return to Innovate UK Business Growth from the inside — mentoring executive teams in high-growth companies, concluding in the AI-readiness feasibility study.

THE TECHNOLOGY

DELIVERED

42 — the Innovate UK Business Growth AI-readiness feasibility study

42 stakeholder perspectives gathered in structured interviews, analysed with a purpose-built retrieval system — five evidenced reports, qualitative and quantitative, with extraction accuracy and full provenance.

DELIVERED

329 — the UK innovation ecosystem

329 stakeholder organisations in a single dataset, every record enriched through structured AI passes. Full coverage, auditable.

CAPABILITY DEMONSTRATION

332 — Horizon Europe

332 funding calls analysed against an organisation’s brief — the matching method, rerun with modern tools. Run as a hypothesis; returned auditable.

DELIVERED WORKING WITH DEEP ECOSYSTEMS

780 — the conversational knowledge platform for Smarterra

780 pages of consortium documentation made structured, navigable and conversational for a cross-national cohort — delivered for a Horizon Europe agri-food ecosystem programme, and presented at its closing summit in Matera, Italy, 2025.

The architect

A potter's hands shaping clay on a wheel

Twenty years inside three sectors established what the structure had to hold: a conversation that reaches real depth, expertise translated into decisions, collaborators matched from evidence, and collaboration opened while what each participant wants to keep private stays private. Horizon Intelligence is built to those requirements — an architecture designed by a practitioner who worked the problem for two decades, and built with an AI team working under that practitioner’s judgement.

Grounded in a research field

Community intelligence is a field of research, with decades of findings behind it — studied at MIT by Alex "Sandy" Pentland and colleagues, and resting on a wider body of work on the conditions under which groups become collectively intelligent: genuine diversity of view, independence of contribution, a bounded and governed group. Horizon Intelligence is a deployable architecture built against those findings. What the field established in the laboratory, the engine makes practical at depth and at scale — a whole network engaged one private conversation at a time, with the aggregate returned in a form the research says collective direction actually requires. The designed architecture is the new capability: this level of engagement was previously impossible to operate, and is now an engine you can run.

The research is cited here as the grounding for the design.

Working with DEEP Ecosystems

The most recent chapter of the record is work delivered alongside DEEP Ecosystems, the European accelerator for innovation ecosystems.

That is the record. What comes next is written with the groups that run their own hubs.

Start with a conversation

A twenty-year vision to close the gap.