RevAIsor secures £800k Innovate UK project to build sovereign “AI Autonomy by design” with SENTINEL Phase 2

An £800,000, 12-month project to mature a sovereign UK capability that proves — not just claims — that autonomous AI agents stay within safe boundaries. Inside SENTINEL Phase 2.

London, UK — RevAIsor has secured an ~£800,000 project from Innovate UK to launch SENTINEL Phase 2: a 12-month initiative to mature a sovereign, fully defensible UK capability for securing autonomous AI systems by design. The award forms part of a three-phase programme under the UK’s £500m Sovereign AI fund. With this new grant, RevAIsor has secured nearly £1m in grants to build the next game-changing technology in AI testing.

From assistive tools to autonomous agents

As AI moves from assistive tools into fully autonomous agents, the risks change shape. Adversarial attacks and unpredictable behaviour escalate fast, and traditional “black box” models make it nearly impossible to formally guarantee what an agent will do — or to defend against novel threats like prompt injection and data exfiltration.

SENTINEL Phase 2 is built to close that gap before deployment, not after.

A mathematical safety shield

At its core, SENTINEL is a deterministic, neuro-symbolic assurance engine — a mathematical safety shield that guarantees AI systems operate strictly within pre-defined, secure boundaries. It pairs a neural System 1 for fast, intuitive perception with a symbolic System 2 for deliberative, logical reasoning: a guardrail that can be inspected, audited and proven.

The goal of Phase 2 is to harden that engine into a sovereign-grade capability the UK can stand behind — “provably safe, deterministic, and interpretable” from the soil to the token.

“This funding from Innovate UK is a massive validation of our vision. We are not just building software; we are building a sovereign AI assurance infrastructure. By embedding security and responsible AI principles at the foundational level, we are enabling enterprises and government bodies to adopt autonomous systems with absolute confidence.”

Dr. Edgar Lopez-Rojas, CEO RevAIsor

Built for finance and defence

Financial services face a verification crisis where black-box models introduce operational and regulatory risk. SENTINEL automates governance to align with the EU AI Act and FCA regulations, and produces deterministic, human-interpretable audit trails — tracing AI decisions back to specific rules to meet requirements like the Consumer Duty.

In defence and critical infrastructure, the stakes of failure are extreme, demanding Safety-by-Design rather than post-deployment auditing. SENTINEL’s symbolic layer formally verifies every action an agent proposes against inviolable, human-defined rules — preventing catastrophic failure modes and keeping sovereign systems resilient against third-party supply-chain vulnerabilities.

What’s next: get involved

With Phase 2 officially underway, RevAIsor is opening three doors:

• 10 enterprise testing partners across defence, finance, healthcare and critical infrastructure, with exclusive pre-deployment access to the SENTINEL assurance framework.
• Strategic investors passionate about DeepTech, AI safety and sovereign defence capabilities.
• Talent — RevAIsor is hiring an AI Engineer (PhD) to help build the core safety architecture.

If you’re interested in becoming a testing partner or joining the current funding round: contact@revaisor.com.