Frequently asked questions
Direct answers to the questions evaluators actually ask — including the ones where the honest answer is 'not yet'.
What does mAIb Tech actually sell?
Enterprise software licences and related services. The flagship product, Locs, is governed private AI infrastructure for telecom operations — installed in your environment under an annual licence, with a fixed-scope paid pilot as the entry point. GAIO serves governance and compliance teams, and AOS-1 is the governance runtime both build on.
Does our data leave our environment?
No — the platform is designed for zero raw-data egress. Models, documents, embeddings, answers and audit records live on infrastructure you control, including fully air-gapped operation. In on-premises deployments you can enforce this at your own network boundary.
Is mAIb Tech certified (ISO 27001, SOC 2, etc.)?
No. mAIb Tech is pre-certification and does not claim an issued certification. The framework alignment page states our per-framework position precisely — designed-to-support, aligned, or roadmap — and it changes only when the underlying evidence changes.
Which AI models does the platform use?
It is model-neutral: open-weight local models by default, or a customer-provided endpoint, behind a model-adapter layer. Models are governed components — evaluated before activation and re-evaluated after changes.
Can the AI take actions in our network?
No. Output is advisory and decision-support only, labelled with confidence and risk indicators. The system refuses to recommend production-changing actions and says so when approved knowledge is insufficient.
How is this different from just running a local LLM?
Local answers where AI runs; it does not answer whether the AI is grounded, permission-aware, evaluated, auditable or reversible. The platform adds governed knowledge ingestion, in-query access control, evaluation against golden sets, tamper-evident audit and rollback — the parts a security review actually asks about.
Do you have reference customers?
We are pre-launch and onboarding a limited design-partner cohort. We do not publish customer names, anonymised silhouettes or testimonials we cannot legally back.
What does the free readiness assessment give us — and what is it not?
It is a 15-question self-assessment producing an immediate score, maturity level and priority gaps — useful for structuring an internal conversation. It is preliminary and non-binding: not an audit, a certification, or legal or regulatory advice.
How do we evaluate the security architecture?
Start with the Security & Trust Centre, then request the security architecture pack: architecture, data flow, identity, encryption, key management, assurance, threat model, evidence, recovery, known limitations and customer responsibilities. NDA available on request.