SecurityBrief UK - Technology news for CISOs & cybersecurity decision-makers
United Kingdom
FIOR AI gateway passes NCSC cyber resilience test

FIOR AI gateway passes NCSC cyber resilience test

Fri, 20th Nov 2026 (Today)
Joseph Gabriel Lagonsin
JOSEPH GABRIEL LAGONSIN News Editor

FIOR said its AI Gateway has completed a Cyber Resilience Testing assessment under a National Cyber Security Centre scheme. The assessment found that all 69 security claims made for the product were met.

The test was carried out by NCC Group, which is assured by the National Cyber Security Centre to conduct Cyber Resilience Testing. FIOR said the product is now described as NCSC CRTF - Cyber Resilience Assessed under the watchdog's Principle Based Assurance scheme.

The result comes as organisations move AI agents from pilot projects into live environments across enterprise applications, data systems, application programming interfaces and infrastructure. That shift has increased pressure on security teams to establish which AI agent is acting, who authorised it and whether an action is within policy before it is carried out.

Cyber Resilience Testing is designed to examine how well a connected product can protect itself against common cyber attacks when connected to the internet. In FIOR's case, the assessment examined its AI Gateway against 18 National Cyber Security Centre principles across five themes: secure design and development; build environment security; secure deployment and maintenance; communication with customers; and product-specific usage, design and operation.

FIOR's product is aimed at organisations that want a separate control layer for AI agents. According to the company, the gateway gives each agent a cryptographically secure identity, determines what that agent is authorised to do and applies those permissions at runtime, while also providing revocation and audit records.

Buyer scrutiny

The announcement comes amid heightened attention on AI-related cyber risk in the UK. Security buyers in government and large companies face a fast-growing market for products intended to govern or constrain AI agents, often with limited third-party assessment to compare competing claims.

By publishing the outcome of an independent review against National Cyber Security Centre principles, FIOR is seeking to give buyers a reference point beyond vendor assertions. It said the full assessment report is available to customers, while a product assessment report is publicly available.

David Williams, Founder and Chief Executive Officer of FIOR, said: "Anyone putting AI agents to work inside sensitive systems needs to be confident that the controls governing those agents are based on an identity that is secure. We believe AI governance cannot be trusted unless the bedrock of identity is cryptographically secure and entirely self-sovereign, with no third-party dependencies. We have made our first customer deployments, but trust and confidence must be earned continually. That is why we chose to put our AI Gateway through an independent assessment using a methodology created and approved by the NCSC. Completing it with no unmet claims gives our government and enterprise customers an objective basis for their decisions and reflects the bar we believe the whole AI agent security market should be held to.

"We are grateful to the NCSC, which has been pragmatic in offering this new programme. Although exhaustive and detailed, we were able to complete it in just two quarters. It forms part of a broader picture in which the UK is becoming a world leader in the assurance of AI models and security. Achieving this status adds further distinction to the FIOR AI Gateway product."

Assessment framework

The National Cyber Security Centre's Cyber Resilience Testing framework is built around risk outcomes rather than a fixed checklist. Assessments are carried out by independent Cyber Resilience Test Facilities assured by the agency.

Under that framework, a completed assessment does not amount to certification or regulatory compliance, and it applies only to the version of the product that was examined. Even so, for suppliers in newer cybersecurity categories linked to AI, an external assessment can carry weight as customers look for more rigorous ways to test product claims.

That is particularly relevant in the market for AI agent security tools, where products are intended to control software agents acting with increasing autonomy across multiple systems. For buyers, the challenge is not only whether those controls exist, but whether they can be independently tested before organisations rely on them in sensitive settings.

FIOR said its gateway is built on post-quantum cryptography and functions like a passport system for AI agents, checking what they are allowed to do before an action takes place. The company said the assessment gives government and enterprise customers an objective basis for procurement decisions.