Guides written by the team that builds it
Practical material for evaluators: architecture deep-dives, security guidance and pilot planning. Articles hosted on the Locs product site open there — same company, same authors.
The architecture of governed private AI
How governed private AI actually works: the appliance model, the eight-step knowledge lifecycle, the model-adapter layer, and runtime trust controls.
Read on locs.maib.io ↗A CISO's guide to on-premises AI
Eight questions security leaders should ask any on-premises AI vendor, the evidence standard to demand, and what a defensible AI approval looks like.
Read on locs.maib.io ↗The security model of governed private AI
The nine governance layers a private AI deployment needs — secret blocking, redaction, two-key approvals, kill switch, tamper-evident audit — each enforced in code.
Read on locs.maib.io ↗The 90-day enterprise AI pilot plan
A phase-by-phase enterprise AI pilot structure that ends in audit-ready evidence and a conversion decision — not an endless proof of concept.
Read on locs.maib.io ↗The GAIO Doctrine (whitepaper)
The published GAIO whitepaper on sovereign AI governance. Citable document — DOI 10.5281/zenodo.20457628, CC BY 4.0.
Read →Security architecture pack
Architecture, data flow, identity, encryption, key management, assurance, threat model, evidence, recovery, known limitations and customer responsibilities.
Available on request →Telecom private AI readiness guide
What readiness for private AI looks like across data control, governance, security and operations — and how to assess your own position.
Read →No fake downloads here: anything marked "available on request" is delivered by a person after a short request — because we version-control what evaluators receive.