Vocabulary Browser
Vocabulary Browser is the reference catalog for the standardized terms used across VeriProof governance data. It helps teams align on intent labels, risk types, agent roles, guardrail categories, and other governed values that power filtering, analytics, and policy logic.
Use it when defining rules, reviewing annotations, or standardizing language across applications and business units.
Vocabulary Browser
The Vocabulary Browser is a searchable catalog of every standardized governance enum value in the VeriProof platform — intent labels, risk types, agent roles, decision outcomes, guardrail actions, and content safety categories. It is the reference point for teams instrumenting their AI systems, authors writing governance rules, and auditors verifying that session annotations use consistent, recognized vocabulary.
All portal roles can read the Vocabulary Browser. Developers and Governance Engineers use it most actively — during SDK instrumentation and rule authoring respectively.
Use the Vocabulary Browser to find the correct enum values before setting governance attributes in your SDK calls. Recognized values from this catalog ensure your sessions filter and aggregate correctly across every portal view.
Rego rules that reference input.veriproof.intent_label or input.veriproof.risk_type must use the exact string values listed here. Use the Copy value button next to any enum to paste it directly into your policy authoring workspace.
Why Governed Vocabulary Matters
When the VeriProof SDK records session annotations, it uses structured enum values rather than free-form strings. Consistent vocabulary is what enables:
- Cross-application aggregation — governance metrics only add up correctly if “loan-decision” means the same thing in Application A and Application B
- Regulatory mapping — framework-to-evidence linking depends on intent labels and risk types mapping to known regulatory categories
- Rule portability — governance rules written against intent labels can be reused across applications without modification
The Vocabulary Browser makes the full catalog visible and searchable so teams can use recognized values during instrumentation rather than introducing ad-hoc labels that fragment your governance data.
Browsing the Catalog
By Category
The catalog is organized into five categories:
| Category | Description |
|---|---|
| Intent labels | Model-classified intent values used to categorize the purpose of an AI interaction (e.g., loan-decision, medical-advice, fraud-detection) |
| Risk types | Structured risk descriptors that complement the LOW–CRITICAL risk level (e.g., bias-risk, data-quality-risk, hallucination-risk) |
| Agent roles | The nine role types assigned to agents in multi-agent workflows (e.g., orchestrator, specialist, validator) |
| Guardrail categories | Standard categories for guardrail rule classification (e.g., safety, compliance, quality, business-logic) |
| Content safety flags | Structured flags for content safety signals (e.g., potential-pii, harmful-content, off-topic) |
By Industry Vertical
Each vocabulary item is tagged with the industry verticals it is most commonly used in. Filter by vertical to see the subset of the catalog most relevant to your use case:
- Banking & Financial Services
- Insurance
- Healthcare & Life Sciences
- Legal & RegTech
- Government & Public Sector
- Manufacturing & Supply Chain
- Telecom & IT Operations
- Higher Education
- Energy & Utilities
- HR & Workforce Management
- GRC & Enterprise Risk
Search
The search bar at the top of the Browser accepts free-text search across item names, descriptions, aliases, and framework tags. Searches are fuzzy — typing loan surfaces loan-decision, loan-underwriting-risk, credit-assessment, and other semantically related terms.
Vocabulary Item Detail
Click any vocabulary item to open its detail page:
| Field | Description |
|---|---|
| Canonical name | The exact string to use in SDK annotations or governance rules |
| Description | What this value means and when to use it |
| Industry tags | The verticals this value is most commonly applied in |
| Framework mappings | Regulatory framework articles or controls this value is relevant to |
| Aliases | Common alternative names this value encompasses |
| Usage count | How many sessions in your tenant currently use this value (your data only) |
| Related items | Other vocabulary items in the same category or with related meaning |
Custom Vocabulary
If your organization uses intent labels or risk types that are not in the standard catalog, you can add custom vocabulary items under Settings → Infrastructure → Enum Catalog. Custom items appear in the Vocabulary Browser alongside standard items, tagged as Custom.
Before adding a custom vocabulary item, search the Browser thoroughly — the catalog is extensive and an existing item often covers what you need. Using existing catalog items preserves compatibility with framework mapping and cross-tenant benchmark features.
Related Documentation
How to use vocabulary items correctly when instrumenting your AI application.
SDK Instrumentation GuideAuthor governance rules that reference intent labels and risk types from the catalog.
Rules BuilderPre-built vocabulary sets and SDK configuration examples for your industry vertical.
Industry Quick Start