Locs — private AI your telecom operations can prove
Locs connects approved telecom knowledge to customer-controlled AI, evaluates every active configuration, blocks unsafe behaviour and measures operational outcomes — all inside your environment.
Software you install. A portal your teams use. Nothing leaves your network.
A software appliance
Enterprise software you run on your own servers — a hardened container package your ops team installs. No cloud service behind it, no account with us: the model, your documents and every answer live entirely on your infrastructure. Runs fully air-gapped if you need it to.
Answers from your own knowledge
Your administrators feed it documents you already trust — SOPs, runbooks, ticket history, RCA reports. Staff ask questions in plain language and get answers with citations, a confidence score and a risk label. If it doesn't know, it says so instead of guessing.
A governed AI workspace
Three telecom copilots (BSS/charging, NOC, customer care), an admin console for approving what the AI may know, an evidence centre your auditors can use, and a signed update mechanism — plus installation, training and a full document kit.
If you've used a public AI chatbot: Locs is that experience rebuilt for a regulated enterprise — inside your network, answering only from approved documents, showing its sources every time, keeping tamper-evident records of everything it does.
Charging & BSS assurance
Charging and billing incidents are long, fragmented and expert-dependent. Locs helps teams retrieve approved knowledge, analyse incidents consistently, identify missing evidence and prepare vendor escalations — advisory only, with refusal of production-changing recommendations.
Why security teams can approve it
Zero egress by design
Raw enterprise data egress is zero by design — enforceable at your network boundary in on-prem and air-gapped deployments.
Evaluated before use
An Assurance Lab measures grounding, citation coverage and refusal recall against versioned golden sets before any configuration activates.
Governed knowledge
Secret blocking, redaction before embedding, human approval and versioned rollback for everything the AI knows.
Enterprise identity
OIDC/Entra integration with role-based access enforced inside retrieval queries.
Tamper-evident audit
Append-only, hash-chained audit trail with SIEM forwarding — honest about its limits: tamper-evident, not WORM.
Kill switch & recovery
A kill switch blocks AI serving regardless of state; encrypted backups with verified restore drills.
From readiness to production, in defensible steps
- Self-assessment — start free with the readiness self-assessment on this site.
- Readiness engagement — a fixed-scope security and architecture review with a deployment recommendation and pilot plan.
- Lighthouse pilot — paid, fixed-scope, on one measured charging/BSS use case in your environment, ending with signed evidence.
- Enterprise licence — annual licence with editions up to fully air-gapped sovereign deployment, plus a Managed Assurance service.
Commercial details are scoped per environment — no invented pricing. Full product site: locs.maib.io