Refractal Intelligence secures £715k seed funding for AI trust layer

Enterprises seeking to deploy AI agents in production environments now have a new trust layer following a £715k seed investment into London-based Refractal Intelligence Limited. The company develops software that intercepts and evaluates every tool call made by an AI agent against specific enterprise policies.

The funding was formalised through an allotment of shares (April 2026) with the corresponding documentation filed the following month (May 2026). This seed round provides the capital for the startup to build out its runtime security infrastructure.

Refractal Intelligence Limited was incorporated in February 2026 to solve the problem of trust and accountability in autonomous AI systems. Its product acts as a gatekeeper, ensuring that AI agents do not exceed their permitted authority or violate compliance rules when interacting with external tools and databases.

A core component of the technology is an append-only audit ledger. This ledger records every action and evaluation, providing a permanent and tamper-proof history of AI agent activity. This feature is specifically designed to enable regulated enterprises to meet audit requirements while using generative AI technologies.

The company is headquartered in London. As Refractal Intelligence Limited is in its early stages of operation, having been founded earlier this year, financial metrics such as annual revenue, total assets, and profit or loss figures are not yet available in public records. The current headcount and specific valuation following this seed round have not been disclosed in the latest filings.