Data requests system blueprint
This system receives business questions through an approved workspace, resolves intent, and answers from governed metrics and datasets. It returns the calculation, definition, source lineage, and relevant limitations, while routing ambiguous or sensitive requests to a qualified owner.
- Business function
- Leadership
- Operating context
- Across the business
- Primary owner
- Head of Analytics
Operating brief
What starts the system
An authorized user submits a supported business question in the connected request channel.
Row, field, and metric access follows the requester's identity
Operating sequence
How the system works
1. Classify the question, requester permissions, intended grain, and required definitions.
2. Query approved data products, validate the result, and attach source and calculation lineage.
3. Return the answer or route ambiguity, restricted data, and novel analysis to an accountable analyst.
Inputs
- Natural-language business question and requester identity
- Governed semantic model, metrics, and data products
- Access policy, definitions, and escalation rules
Outputs
- Sourced answer with definitions and calculation context
- Escalation or reusable-query record when human analysis is required
Controls
- Row, field, and metric access follows the requester's identity
- Unsupported inference and low-confidence interpretation require human review
Source landscape
Examples of systems it may connect to
- Slack
- Microsoft Teams
- Snowflake
- Google BigQuery
- Microsoft Power BI
Reusable core
What can carry across companies
The reusable core is intent resolution, governed query generation, permission enforcement, lineage, and escalation.
Your installation
What must be adapted
Supported questions, semantic definitions, channels, access groups, answer formats, and confidence thresholds are local.
Start with the closest pattern–or start from your process.
The catalogue represents reusable operating shapes, not fixed off-the-shelf applications.