The consistent context your AI needs
CData gives teams a governed layer so every prompt and workflow operate from the same contextual understanding.
The Problem
When business context lives in prompts and raw tables, answers drift fast.
Without a governed definitions, every agent and dashboard builds its own version of ""Q2 revenue"" or ""active customers."" The result isn't just inconsistency—it's loss of trust.
When AI builds logic from scratch against raw data, every interaction is another chance to get it wrong. Ttool calls, variance, and token costs compound with every agent you deploy.
Without scoped workspaces, every new agent inherits access it doesn't need. That turns every use case into a manual access-design project.
How It Works
One governed context layer for definitions, access, and performance.
Define logic once, scope who can use it, and tune how fresh it stays—one control plane behind every agent, analyst, and dashboard.
Data teams define trusted business context schema once as a derived view, such as “Q2 revenue by region” with the right fiscal calendar, date field, and revenue definition.
That derived view is registered in the platform and surfaced automatically in getSchema, so AI discovers and uses the governed definition instead of building the query from raw data.
Each agent or team operates inside a workspace scoped to only the data sources, tools, and derived views it is authorized to access.
Query cache and scheduled refreshes are configured per view for high-frequency metrics, using customer-controlled storage and the right freshness cadence for each use case.
Every interaction is logged under the same governance model, whether the agent or user is querying a live source, a governed view, or a cached definition.
One governed definition means one governed answer. Everywhere.
Key Capabilities
Consistency, isolation, and performance built into the context layer.
Derived views
Trusted business definitions live in one place
Define revenue logic, pipeline views, fiscal-period calculations, and cross-system metrics once as reusable derived views.
Query cache
High-frequency queries stay fast without hammering source systems
Use query cache for large, frequently requested datasets while keeping storage customer-controlled.
Workspaces
Each agent works inside a governed workspace
Scope every agent to the specific data sources, tools, and definitions it is allowed to use, and nothing more.
Freshness
Freshness is tunable per business definition
Set refresh cadence independently for each governed view based on how often the underlying data actually changes.
Replication API
Embedded teams can control sync behavior programmatically
Use the replication API to manage sync frequency, incremental strategies, and data subsetting across customer environments.
Schema discovery
AI discovers governed views before it builds from scratch
Derived views surface automatically in getSchema so agents use the trusted definition instead of reconstructing business context ad hoc.
Meeting customers where their data lives
Built for how teams actually work with data today.
One governed definition means one governed answer. The same business question returns the same result everywhere it is used.
VP of Business Intelligence Platforms
Implementation Path
Configure context control in week 1
Define core views
Connect priority data sources and define the first derived views for the highest-value business metrics.
Milestone: Critical metrics such as "Q2 revenue by region" are available as governed definitions from the first interaction.
Scope workspaces
Configure workspaces for active agent and team use cases, with scoped data access and tool exposure.
Milestone: Each agent operates inside its own authorized boundary instead of sharing platform-wide visibility.
Tune performance
Configure cache, refresh cadence, and replication controls for production workloads.
Milestone: Governed definitions are serving consistent answers at production-ready performance levels.
Security & compliance
Data control that preserves governance.
- Passthrough identity on derived views — user-level permissions enforced on every query.
- Customer-controlled cache — query cache stored in your own Postgres instance.
- Workspace isolation — each agent scoped to authorized data and tools only.
- Query-level audit logging — every derived view query logged with full metadata.
- No data movement by default — live queries go to source, and cache is opt-in.
- SOC 2 Type II — Completed.
- ISO/IEC 27001:2022 — Completed.
- GDPR support — live access plus customer-controlled storage options for governed data paths
FAQ
Questions teams ask first.
One definition of your business logic. Consistent answers everywhere it gets used.
Talk to our team about defining your first governed views and scoping your agent workspaces. Or explore how the Data Control Plane fits into the broader CData platform.