Colchix builds an operating system for safe enterprise AI adoption

The rapid adoption of generative tools across modern corporate teams creates invisible security risks for information technology leaders. Employees routinely input proprietary code, financial records, and private customer details into browser tools, desktop apps, and coding plugins. Italian technology entrepreneur Gianni Valerio and technical co-founder Gabriele Salvo recognised this visibility gap when platforms like ChatGPT, Claude, and Gemini spread rapidly across enterprise networks. Based in Amsterdam, their venture Colchix B.V. designed a unified console to monitor all enterprise tools, accounts, and internal agents. The co-founders aimed to let organisations adopt workplace tools at maximum speed while remaining completely secure and compliant.

To solve the visibility problem, the founders created the Colchix AI Governance OS, a dedicated security suite that operates without adding operational friction. The software monitors ten distinct pathways through which software enters an enterprise, including web chat interfaces, APIs, model context protocol servers, and embedded software tools. Rather than relying on manual reporting, the console provides real-time visibility across all finance, engineering, and support teams. Hosted entirely within the European Union, the infrastructure never stores user prompts, allowing regulated enterprises to retain full sovereignty over their private data flows.

The platform executes its core governance tasks through three specialised autonomous software agents named ARGUS, TALOS, and ATHENA. The ARGUS agent continuously discovers all software usage across an organisation, categorising applications into approved tools, shadow software, and personal accounts. Before sensitive information reaches external models, the TALOS agent automatically masks confidential data values and securely restores them inside the returned response. Meanwhile, the ATHENA agent monitors daily digital interactions, turning ongoing employee activity into structured, audit-ready documentation for internal risk teams.

Underpinning these automated operations is THEMIS, a proprietary System One² model built by Colchix to process governance requests in milliseconds. By constantly evaluating network traffic, THEMIS enables ATHENA to produce definitive audit evidence for major regulatory frameworks, including the EU AI Act, GDPR, and ISO/IEC 42001. This continuous verification process allows compliance teams to satisfy complex legal mandates without creating tedious administrative work for employees. As a result, enterprises can easily track both human staff and autonomous software agents within a single clear dashboard.

To ensure early-stage ventures can also close corporate sales, the team launched the Colchix for Startups programme. Qualifying startups under fifty people gain access to the complete agent suite, allowing them to provide clear security evidence during enterprise vendor evaluations. Looking ahead, the Amsterdam-based team plans to extend its protection layer across emerging protocols, extensions, and local models. The founders will present their AI Governance OS at Web Summit in Lisbon, demonstrating how modern organisations can safely deploy intelligent systems at scale.

Find out more at colchix.com