Agentic AI for UK
government, defence, and national security.
New Voyager builds production-grade agentic AI systems for public sector deployment and conducts frontier AI research for defence and national security applications. We are active participants in ARIA, DSIT, DASA, and Innovate UK innovation programmes.
UK government
AI programmes
we participate in.
ARIA — Scaling Trust Programme
New Voyager is engaged with ARIA's Scaling Trust programme, which funds the development of open-source tools for secure multi-agent coordination. Our work on inter-agent verification and human-in-the-loop oversight aligns with this programme's Track 2 (Tooling) objectives. Backed by nearly £50m, this programme seeks to create the capability for AI agents to securely coordinate, negotiate, and verify with one another.
DSIT — GOV.UK Agentic AI Companion
The Department for Science, Innovation and Technology is procuring agentic AI systems to power GOV.UK citizen services. New Voyager's multi-agent architecture — built for action-taking, context retention, and seamless system integration — aligns with DSIT's published procurement requirements for sophisticated agentic AI.
DASA — Defence & Security Accelerator
The Defence and Security Accelerator funds innovations that strengthen UK national security. New Voyager's agentic AI systems address DASA's priority areas directly: secure multi-agent coordination across classified environments, multi-source intelligence fusion pipelines that synthesise disparate data streams in real time, edge-deployed autonomous agents for disconnected or contested operational theatres, and adversarially robust decision-support systems with full audit provenance. Our architecture is built for environments where latency, trust, and chain-of-custody are non-negotiable.
Innovate UK — AI Champions (Frontier AI)
Innovate UK's AI Champions programme funds UK SMEs developing frontier AI and machine learning technologies. New Voyager's research into novel multi-agent architectures, secure coordination protocols, and explainable autonomous decision systems aligns with Theme 3 (Secure AI for National Security & Defence) — specifically multi-source fusion analytics, edge autonomy, and adversarially robust AI at low technology readiness levels.
AI capabilities built for national security contexts.
Purpose-built agentic architectures for environments where security classification, adversarial threat, and operational tempo define the engineering constraints.
Secure multi-agent coordination
Agent systems designed for classified and sensitive environments — with inter-agent verification, cryptographic audit trails, and zero-trust communication protocols between autonomous subsystems.
Multi-source intelligence fusion
Distributed agent pipelines that ingest, correlate, and synthesise data from heterogeneous sources in real time — producing structured analytical outputs with full provenance tracking and confidence scoring.
Edge autonomy & disconnected ops
Lightweight agent architectures deployable at the tactical edge — capable of autonomous operation in disconnected, intermittent, or limited-bandwidth environments with graceful state reconciliation on reconnect.
Adversarial robustness & explainability
AI systems hardened against adversarial manipulation with interpretable decision pathways. Every agent output includes structured reasoning chains and confidence bounds — critical for human-on-the-loop operational contexts.
Why specialist agentic AI matters for government and defence.
01
Compliance by design
Every agent system we build includes human oversight gates, full audit trails, and configurable approval workflows — meeting the responsible AI deployment standards required for public sector use.
02
Proven multi-agent architecture
Our systems use LangGraph-based orchestration with defined agent roles, inter-agent messaging protocols, and state management — not off-the-shelf chatbot wrappers. This level of architectural rigour is what government AI procurement evaluators require.
03
SME specialist advantage
As a specialist SME we are eligible for DASA Innovation Loans, Innovate UK AI Champions grants, and ARIA programme funding. We operate at the intersection of deep technical capability and procurement agility — moving faster than large system integrators without sacrificing engineering quality.
04
Frontier research capability
Beyond production deployment, we conduct novel AI research at TRL 2-3 — including secure coordination protocols, adversarial robustness methodologies, and explainable multi-agent decision architectures. This dual capability makes us a credible partner for both operational delivery and early-stage innovation programmes.
Working on a government or defence AI programme?
Whether you are a government body scoping an agentic AI deployment, a defence prime looking for a specialist subcontractor, a fellow SME seeking a technical partner for a joint bid, or preparing an Innovate UK AI Champions application — we would like to hear from you.