For consumer use, that trade-off is often acceptable. For organisations with serious data-handling responsibilities — professional duty, client privilege, commercial sensitivity — it is not.

Ava exists to give enterprises a path to AI that does not require giving up control of their data, infrastructure, or outcomes. The work is split between consulting — adoption, governance, and training inside large organisations — and engineering — local, on-device, and self-hosted deployments.

The two halves inform each other. Consulting keeps the engineering grounded in real operational problems. The engineering means clients get architecture built for the long term, not patched together per engagement.

Active client engagement. Currently delivering an AI Adoption Programme for a national transport operator, inside a Microsoft 365 estate.

Innovate UK partner. Technology partner on an Innovate UK funded research project in applied AI.

Open-source contributor. Core contributors to Nano Collective, a community building local-first AI tooling. Its flagship project, Nanocoder, has 2,000+ GitHub stars.