Every company already has a data control plane. It’s a person.
The senior data engineer who knows where the customer table actually lives, which of four copies is current, and who’s allowed to see it. That worked when the askers were people. It breaks when the askers are agents. They can’t walk over to someone’s desk.
Today that layer is a person and a ticket.
AI moves at question speed. Data engineering moves at ticket speed.
The institutional knowledge required to maintain a mature data stack — why these joins, why these exclusions, why this scheduling logic — lives in the heads of two or three engineers. When they leave, the organization regresses. Critical context lives in Slack threads, engineers’ heads, and stale Confluence docs. Onboarding takes weeks.
Tribal knowledge, turned into machine-readable context.
Dagen connects to the catalogs and runtimes you already own, reads tables, views, pipelines, and legacy ETL, and puts what the senior data engineer knows onto the map. Tribal knowledge becomes a map agents and humans can share.
Definitions
Agent Views carry schema, business definitions, and governance metadata — not cryptic columns and missing definitions.
Lineage
Live lineage and dependencies, traced from a field or table back through its transformations to the original source.
Access policy
Policy travels with the data. Row- and column-level permissions apply before an agent ever sees a record. What a person can’t see, their agent can’t either.
Which copy is current
Agents query the view — not the raw warehouse — with definitions, lineage, and freshness attached, and see the last-refreshed time for each view.
How pipelines actually behave
Freshness, last run status, and any open remediations for the pipelines feeding an Agent View.
Memory that compounds with use.
The tribal knowledge that typically walks out the door when a senior engineer leaves is instead captured, structured, and made available to every future agent and engineer who touches the system.
Agent Views over MCP.
Agent Views sit on the map — governed, machine-readable surfaces with definitions, access, and lineage attached. They are exposed through the Dagen MCP server so Claude, Copilot, or any MCP-compatible client can discover, describe, and query them safely. That is how Dagen serves context to agents at runtime through MCP, instead of through a person and a ticket.
The same four tools documented on the MCP reference — list, query, lineage, and pipeline status. See the MCP hub for governed enterprise data over MCP.
list_agent_views
Discover which Agent Views the agent is authorized to reach.
query_agent_view
Query a governed view; definitions and access apply automatically.
get_lineage
Return the lineage graph from source to Agent View.
get_pipeline_status
Report health of the pipelines feeding an Agent View.
A control plane is not a catalog or an MCP gateway
A catalog describes assets for humans to browse. A data control plane for AI agents makes the estate queryable under policy: Agent Views over MCP return rows with governance and lineage, not just documentation.
| What it is | Where it stops | |
|---|---|---|
| Catalog or wiki | Describes assets for humans to browse | Documentation, not governed answers an agent can query |
| MCP gateway | Gateway and connectors in front of sources | Integration surface — not Agent Views with lineage and policy attached |
| Dagen | Cross-estate map + lineage + Agent Views over MCP; pipelines generated onto the engines you already run | — |
Control plane ≠ catalog or MCP gateway. Dagen sits above what you already run and makes that estate usable by agents — without asking you to rip and replace the catalog, warehouse, or orchestrator.
Tribal knowledge and AI agents, answered.
How do we stop depending on one senior data engineer for tribal knowledge when agents need data?+
Today that knowledge lives in a person and a ticket. When that engineer is unavailable — or leaves — agents guess. Dagen captures definitions, lineage, access, and how pipelines actually behave into the map and into Agent Views, so tribal knowledge becomes structured context every agent and engineer can use. Episodic, procedural, and institutional memory compound with use instead of walking out the door.
What happens to tribal knowledge when a senior engineer leaves?+
The tribal knowledge that typically walks out the door when a senior engineer leaves is instead captured, structured, and made available to every future agent and engineer who touches the system. Every pipeline decision, remediation, and pattern is captured in a persistent knowledge base — available to every future agent and new hire.
How do AI agents get that context at runtime?+
Dagen feeds governed enterprise data to AI agents over MCP through Agent Views on the estate map. Agents call tools such as list_agent_views, query_agent_view, get_lineage, and get_pipeline_status — documented on the MCP reference. That is governed MCP: answers that are allowed, fresh, and traceable — not a raw gateway to tables.
Is a data control plane for AI agents the same as a data catalog?+
No. A catalog describes assets for humans to browse. A data control plane for AI agents makes the estate queryable under policy: Agent Views over MCP return rows with governance and lineage, not just documentation. Dagen is the data control plane for AI.
Do we have to migrate off our warehouse to use Dagen?+
No. Reuse over rip-and-replace. Open table formats and AI-in-front-of-data are why pipelines need a control plane; they are not a mandate to abandon Snowflake or Databricks. Many teams adopt Iceberg on object storage (S3 on AWS) while keeping the warehouse for the workloads that belong there. Dagen agents carry forward the logic in your existing SQL, models, and jobs.
Explore the control plane
Data control plane
Map the estate. Govern access. Serve Agent Views over MCP.
ProductAgent Views
The governed interface between your data and every AI agent.
ProductMCP
Governed enterprise data for AI agents over MCP — the hub.
ReferenceMCP reference
Tools and resources the Dagen MCP server exposes.
Ready to take control of your enterprise data?
Map the estate. Govern access. Serve Agent Views over MCP — from the data you already run.