KnowledgeBricks replaces zero systems. It runs on top of your existing identity provider and data sources with a fully interchangeable AI layer. Every integration is production-grade, documented, and enforced server-side — not wired together with brittle webhooks.
KnowledgeBricks plugs into your existing identity infrastructure without replacing it. Out of the box, it supports email/password, Google OAuth, Microsoft OAuth, and magic-link sign-in. For enterprise deployments, SAML 2.0 SSO connects to any standards-compliant identity provider — Okta, Azure AD, Google Workspace, and others — with no proprietary configuration required.
Access tier assignment is enforced on every server-side request through your existing identity session. Role changes take effect immediately — no cache invalidation lag, no eventual consistency gaps. An admin changing a user's access tier sees that change enforced on the next query, not the next login.
Custom portals ingest your organization's internal knowledge through the same structured ingestion process as the standard portals, with additional data connectors for enterprise document sources. SOPs from SharePoint, operational runbooks from Confluence, bid templates from Google Drive, master data from your ERP, all transformed into queryable knowledge.
Ingestion is not a one-time import. We schedule recurring sync jobs against your sources so the knowledge base reflects your current operations, not a snapshot from 18 months ago. New documents are automatically classified, chunked, embedded, and made available to your team without manual intervention.
SAML 2.0 SSO for enterprise teams connects to any standards-compliant identity provider. Google and Microsoft OAuth for team deployments. MFA available on all plans. Zero passwords stored in KnowledgeBricks infrastructure.
Every retrieval call, role change, and administrative action is logged server-side with a full audit trail. Logs are exportable and available to your security team on demand. No action inside KnowledgeBricks is unrecorded.
The retrieval and generation layer is provider-agnostic. Embedding models and answer-generation models are independently configurable per portal. Swap providers without re-architecting the platform. All LLM calls are logged and auditable.
Relational data and vector embeddings are handled within a single SOC 2 Type II-certified infrastructure (Supabase). Row-level security enforces portal data isolation at the database layer — no cross-tenant data access at any level.
Query patterns, session depth, feature usage, and content gap signals are tracked to surface maintenance priorities and usage insights. All analytics data is anonymized. No PII in event payloads.
Account activation, password resets, and admin alerts delivered via production-grade transactional email. Custom domain sending available for white-label custom portal deployments.
Versioned REST APIs expose engagements, projects, and deliverables, and signed webhook events fire on status changes and milestones — push updates into Monday.com, Zapier, or n8n without leaving your workflow tools.
OAuth-connected integration pulls source documents from SharePoint and OneDrive and exports proposals and status updates straight back, with encrypted token storage and automatic refresh.
Custom portal builds start with a data inventory call. We identify your highest-value knowledge sources, assess document quality and format, and design a collection architecture that pulls the right content without granting over-broad access to your systems.
We collect SOPs, runbooks, training materials, bid templates, and methodology documents. Any format, PDF, Word, Excel, HTML, Markdown. Legacy scanned documents handled via OCR pipeline.
ERP item masters, WMS rate cards, TMS lane data, and supplier databases are exported via API or scheduled flat-file. Structured data is formatted into queryable knowledge entries, not raw database rows.
Confluence, Notion, and SharePoint wiki spaces are connected via API with read-only service accounts. Sync schedules are configurable, daily refresh for stable reference content, hourly for active operational data.
The knowledge base is not a one-time import. We maintain ingestion pipelines, retire outdated content when sources are updated, and flag knowledge gaps identified by user query patterns.
A scoping call covers your existing stack, identifies integration points, and produces a data architecture plan for your custom portal, or confirms a standard portal fits your needs out of the box.
Standard portal integrations are pre-built. Custom portal data connectors scoped during onboarding.