AI Score raises $5.4M to keep enterprise AI agents on a leash

Why it matters

AI agents turn governance into a runtime problem: enterprises need to know what software can access and do while it is acting. AI Score's round funds that control layer just as regulators and standards bodies begin demanding evidence.

Alex Harland and Benita Tibb have raised $5.4 million for AI Score, betting that enterprises will need continuous oversight of AI agents rather than another collection of policies and annual audits.

AI Score said Friday that Fuel Ventures led the seed round, with founding investor GALLOS Technologies participating. The financing, reported by Tech.eu, also included Alan Morgan, the co-founder of MMC Ventures, and Mo El Husseiny, managing partner of Ventura Capital.

AI Score was co-founded and is led by Alex Harland, a former National Cyber Security Centre executive who previously held corporate cybersecurity roles at Compare the Market and BGL Insurance, according to Sky News. Tibb, AI Score's chief operating officer, is a former City lawyer.

That pairing explains AI Score's pitch. Harland brings experience responding when systems and controls fail under pressure. Tibb brings the legal and operational view of how accountability is assigned inside large organisations. They founded AI Score around the contention that AI governance must follow software continuously, including after autonomous agents gain access to corporate tools and data.

"AI operates at a scale and level of automation that traditional governance simply wasn't built for," Harland said in the funding report.

From policy documents to runtime controls

AI Score says its API-first platform connects to enterprise AI services, cloud applications and self-hosted agents. It maintains an inventory of AI assets, maps activity to company policies and regulatory frameworks, and monitors how agents interact with tools, permissions and workflows.

For AI agents, the difficult part begins after deployment. An agent may receive permission to read files, query a database, send messages or trigger another application. Its actions can change according to prompts, context and the output of other models. A policy approved months earlier gives a board little visibility into what the agent did on Tuesday morning.

AI Score says it addresses that gap through live monitoring, automated controls and audit records. The product is offered as private-cloud software and uses gateways and direct integrations to extend coverage across a customer's AI systems. Claims that it can be deployed within minutes and provide a complete view of an organisation's AI estate remain AI Score's own; AI Score has not published independent performance tests.

The approach reflects Harland and Tibb's central thesis. On its company page, AI Score describes governance as an operating capability that should help enterprises move faster while keeping risk, cost and accountability visible. That framing matters commercially. Security software sold solely as a brake on AI adoption faces a harder budget conversation than software presented as the control layer that lets a company deploy more AI.

AI Score points to Clifford Chance and professional-services group S&W as users or partners. AI Score's website displays testimonials from Anthony Vigneron, director of legal technology solutions at Clifford Chance, and Mark Hendry, a partner at S&W, while describing its broader customer base only by sectors such as legal services, financial services and consumer brands.

A young company with an unusually senior bench

Companies House records show AI Score was incorporated on April 11th, 2025. AI Score emerged from stealth on November 10th, 2025, with more than GBP750,000 in pre-seed financing after being built inside the GALLOS venture studio. The current round comes less than 10 months after that launch.

Harland and Tibb have surrounded AI Score with advisers whose careers sit close to the buyers they need to convince. AI Score lists former GCHQ director Sir Jeremy Fleming, former HSBC Bank Europe CEO Colin Bell and Darktrace co-founder Nick Trim among its advisers. Jonathan Kewley, co-chair of Clifford Chance's global technology group, and former UK national-security official Josh Burch serve as directors.

The group gives AI Score access to boardrooms where operational risk, legal exposure and cybersecurity already carry dedicated budgets. It also raises the standard AI Score will be expected to meet. Enterprises buying governance software will ask for evidence that the controls work across their existing AI services, identity systems and security tools. Adviser credentials can open those conversations; product coverage and deployment results will determine whether AI Score keeps them.

Regulation helps create the budget

The timing gives Fuel Ventures a clear regulatory tailwind. The European Commission began enforcing significant portions of the EU AI Act on August 2nd, 2026, including transparency requirements for certain AI systems. Some high-risk provisions have later application dates, but companies operating in Europe already face growing demands to document how AI is deployed and governed.

The United States is moving through standards rather than an EU-style horizontal law. In February, the National Institute of Standards and Technology launched an AI Agent Standards Initiative focused on interoperability, security, identity and authorisation. Both tracks increase the value of accurate inventories and records of what agents can access and what actions they take.

Investors have already placed larger bets around that control point. Mountain View, California-based WitnessAI announced $58 million in strategic funding on January 13, 2026 for an AI security platform that monitors agent activity, tool access and runtime commands. Willow, an enterprise AI agent governance startup, announced a $7 million seed round on June 4, 2026 for technology that gives agents identities, scoped access, runtime guardrails and audit trails tied to a human.

AI Score is entering that contest with a broader governance pitch spanning policy, risk, usage, cost and business value. The $5.4 million will fund product development and go-to-market expansion, with a particular focus on agent controls.

Harland and Tibb's wager is that enterprises will eventually treat AI governance as live infrastructure, much as they treat identity, cybersecurity and cloud monitoring. The number of companies making the same wager means AI Score will need to prove its intelligence layer produces decisions and interventions that narrower security tools cannot. The founders now have the capital and regulatory timing to run that test inside more enterprises.