Governed source registry
Track source name, kind, version, URI, citation label, and descriptive metadata for each retrieval asset that supports an AI system.
Feature · RAG sources
Govern the knowledge sources your RAG systems depend on
SentinelAI gives teams a governed source registry for files, URLs, policies, FAQs, runbooks, and other retrieval inputs with versioning, citation context, ingestion status, and AI-system linkage.
What this area covers
RAG source workflows help teams preserve a structured record of the governed content behind retrieval-augmented AI systems. Instead of treating source collections as opaque vector-store inputs, SentinelAI keeps source metadata, version history, chunking signals, and system dependencies reviewable.
Related product areas
- AI systems
Track governed runtime systems that combine models, approved use cases, datasets, release state, and readiness into one operational record.
- Prompt registry
Govern versioned prompts, retrieval settings, linked AI systems, and evaluation posture from a dedicated prompt operations record.
- Dataset governance
Bring datasets, lineage, approvals, taxonomy-backed controls, catalog integrations, and quality gates into the AI governance workflow.
- Release governance
Manage AI-system release records with approval state, rollback references, dependency snapshots, and invalidation handling.
- Evaluation suites
Define governed prompt evaluation suites with baselines, regression thresholds, run evidence, and release-blocking posture.
- Semantic governance
Operate taxonomy, ontology, relationship, and graph-backed governance workflows across models, use cases, datasets, controls, and evidence.
Core capabilities
Built to support production governance work
Ingestion and activation state
See whether a source is draft, ingesting, active, error, or archived so teams can understand which knowledge assets are usable right now.
Chunk and citation visibility
Preserve chunk counts and citation-oriented context so reviewers can understand how governed knowledge becomes retrievable evidence.
AI-system linkage
Connect each source to the AI systems that rely on it instead of leaving retrieval dependencies implicit or undocumented.
Version-aware source maintenance
Create a durable record for source updates and retirements so retrieval changes can be reviewed alongside release and prompt changes.
Target users
- Application and prompt teams governing retrieval sources behind production assistants and copilots
- Compliance and risk stakeholders reviewing knowledge provenance and source appropriateness
- Platform and ML teams managing ingestion posture and system dependencies
- Audit and assurance reviewers who need clearer traceability from generated output back to governed source material
Governance value
- Brings retrieval content into the same governed inventory as prompts, AI systems, and releases
- Improves traceability for source changes that can alter AI behavior after deployment
- Gives reviewers clearer visibility into knowledge provenance and citation context
- Supports safer release and evaluation workflows for RAG-backed systems
- Reduces reliance on undocumented source collections and ad hoc retrieval configuration
How teams use it
A practical operating flow for this feature family
Step 1
Register and classify sources
Capture the type, version, URI, citation labeling, and metadata for each governed retrieval source entering the workflow.
Step 2
Track ingestion readiness
Monitor whether a source is ingesting cleanly, active for use, or needs remediation before it can support governed retrieval.
Step 3
Link into prompts and releases
Use the source record as the knowledge-layer reference when prompts change, releases move forward, or incidents require source review.
Continue exploring
Explore how SentinelAI connects adjacent governance workflows
AI systems
Track governed runtime systems that combine models, approved use cases, datasets, release state, and readiness into one operational record.
Prompt registry
Govern versioned prompts, retrieval settings, linked AI systems, and evaluation posture from a dedicated prompt operations record.
Dataset governance
Bring datasets, lineage, approvals, taxonomy-backed controls, catalog integrations, and quality gates into the AI governance workflow.
Release governance
Manage AI-system release records with approval state, rollback references, dependency snapshots, and invalidation handling.
Evaluation suites
Define governed prompt evaluation suites with baselines, regression thresholds, run evidence, and release-blocking posture.
Semantic governance
Operate taxonomy, ontology, relationship, and graph-backed governance workflows across models, use cases, datasets, controls, and evidence.