Secure Knowledge Base AI
A secure knowledge base AI built for regulated enterprises.
Permission-aware retrieval. No training on your data. Swiss-hosted. Full audit trail on every answer.
If a chatbot can leak a contract, summarize a salary file, or expose a customer record, it's a liability, not a product. Uthereal's secure knowledge base AI is engineered for environments where the answer to 'who can see what?' has to be exactly right — every query, every user, every time.
What you get
Permission-aware
Inherits source ACLs. Users see only what they're authorized to see.
Zero data training
Your knowledge, prompts, and conversations never train shared models.
Cited and auditable
Every answer ships with sources and a full retrieval trace.
How it works
- 1
Ingest securely
PII redaction, classification, and per-source policy enforcement at ingestion time.
- 2
Retrieve under policy
ACL-aware retrieval, sensitive-topic guardrails, and configurable refusal logic.
- 3
Generate with proof
Grounded responses with citations, plus exportable evidence for auditors.
Use cases
- HR and legal knowledge with strict access boundaries
- Healthcare clinical reference under FOPH
- Financial services knowledge under FINMA
- Government and public sector classified content
- M&A data room assistants
- Engineering knowledge with IP separation by team
Built on Swiss-grade trust
Swiss-hosted
Hosted in Switzerland under FADP/nFADP. EU/US regions available.
Enterprise identity
SSO/SAML, OIDC, SCIM, RBAC, IP allowlists, optional CMK.
Compliance-ready
Aligned with ISO 27001, SOC 2, and the EU AI Act. DPA available.
Frequently asked
What makes a knowledge base AI 'secure'?
Three things: data isolation (dedicated tenant, no training on your content), permission-aware retrieval (users only see what they're allowed to), and verifiability (citations, audit logs, evaluation). Uthereal ships all three by default.
Will users ever see content they shouldn't?
No. We inherit document-level ACLs from source systems and enforce them at retrieval time. If a user can't open the source, the AI won't answer from it.
How is sensitive data handled?
PII and secret redaction at ingest, configurable retention windows, optional customer-managed keys, and DLP integrations. Encryption in transit and at rest by default.
Where is the knowledge stored?
Swiss data centers under FADP/nFADP by default. EU and US regions available. BYOC into your own cloud for highly regulated deployments.
Can we prove compliance to auditors?
Yes. Every answer can be traced back to the documents retrieved and the policy in force. Exportable audit logs and evidence packages support ISO 27001, SOC 2, and EU AI Act reviews.
Explore more
A knowledge base AI your CISO will sign off on.
Book a security-first demo. We'll show ACL enforcement, audit traces, and Swiss residency on day one.