AI Score raises £4M seed to keep enterprise AI agents in check

London’s AI Score has raised a £4M seed round led by Fuel Ventures, with founding investor GALLOS Technologies returning, to scale its real-time governance platform for the generative and agentic AI now spreading through large enterprises.

Every large company in Europe is deploying AI agents faster than anyone can supervise them – and that gap between deployment and oversight is precisely where AI Score has planted its flag. Below, I lay out the round, the team behind it, and why a £4M cheque into governance tooling says more about where enterprise AI is heading than most nine-figure model rounds.

A £4M seed led by Fuel Ventures

AI Score announced the £4M round (around €4.6M, per EU-Startups) on 4 September 2026. London’s Fuel Ventures led, and GALLOS Technologies – the security and defence venture studio that incubated the company – re-invested alongside a pointed list of angels: Alan Morgan, co-founder of MMC Ventures; Mo El Husseiny, managing partner at Ventura Capital; and Robert Mann, founding partner at Marshall Bridge Ventures.

The money goes to accelerating the platform and expanding go-to-market, with what the company describes as a particular focus on scaling its tools for managing agentic AI across enterprises. It lands ten months after a £761k pre-seed from GALLOS in November 2025 (Startup Mag) – a quick turnaround that suggests the demand arrived before the sales team did.

“AI operates at a scale and level of automation that traditional governance simply wasn’t built for,” said co-founder and CEO Alex Harland in the announcement. “This investment allows us to accelerate our ambition to reinvent governance for the AI age.”

From the NCSC to the governance layer

Founded in 2025 inside GALLOS’ London studio, AI Score is what happens when national-security people go commercial. Harland was part of the founding team at the UK’s National Cyber Security Centre; co-founder Benita Tibb is a former City lawyer; tech-law specialist Jonathan Kewley co-founded the company and advises it. The advisory board includes Sir Jeremy Fleming, the former head of GCHQ. The team stands at 26 people (Startup Mag, September 2026).

The product gives an enterprise near real-time visibility of its AI estate: which models and agents are running, who uses them, what they cost, how they behave – with audit trails and behavioural guardrails for agents layered on top. The customer list is already the tell: Magic Circle law firms, a FTSE 250 company, a leading UK fintech and global consumer brands (EU-Startups, September 2026). Regulated, reputation-heavy buyers – exactly the organisations that cannot let a fleet of autonomous agents run unsupervised.

A $0.89bn market compounding at 45.3%

The AI governance market was worth $0.89bn in 2024 and is forecast to reach $5.78bn by 2029, a 45.3% CAGR (MarketsandMarkets, August 2026). Few software categories compound at that rate; fewer still are being pulled forward by regulators rather than pushed by vendors.

For scale, look at our own records from the same week: Atira’s $15M seed for AI in industrial sales engineering, and Ki 13’s $5M seed for synthetic fuels. Against those, £4M is a standard European seed – but in a category growing at 45% a year, a standard seed buys an outsized head start. The good news is that a market this young has no incumbent to dislodge: the platforms being picked now become the default audit trail later.

Governance is becoming the buying trigger

Here is the pattern worth watching. The US ships the agents; we write the rulebook – the EU AI Act’s obligations keep phasing in through 2026 and 2027, and even the UK’s lighter-touch approach leaves boards personally uncomfortable with unaudited AI. That tension used to be filed under European roadblocks. It is starting to look like a European market instead.

So where does a 26-person London startup fit in? AI Score’s cap table is the answer: a security studio, a fintech-heavy seed fund and angels drawn from technology, finance and national security (Startup Mag). This is the same playbook that turned Israel’s military-intelligence alumni into a cyber industry – state-grade security talent commercialising into enterprise software. Britain’s NCSC and GCHQ alumni are now doing it with AI oversight, and the buyers signing up first are the most regulated names in the economy.

What we are seeing across the rounds we track is the emphasis shifting from building models to trusting them – and the trust layer is one Europe can plausibly own. Watch whether governance platforms become the system of record for enterprise AI, the way observability became the system of record for cloud. If they do, this £4M will look like one of the cheaper tickets into that race. The opportunity is clear.