Outborn Agent Platform · Step-by-step guide
Data & Business Intelligence
Prepare governed structured data for AI analysis, semantic context, dashboards, reports, and AI Analyst workflows.
Before you begin
- Identify an approved structured source and its data owner.
- Prepare read access limited to the intended dataset.
Follow the workflow
- 01
Create a Data Collection
Add the permitted source and select the schemas, tables, files, or resources the workflow needs.
- 02
Describe business meaning
Define important metrics and dimensions where the source supports semantic metadata.
- 03
Grant analyst access
Make the collection available to the intended Space and analyst or agent, keeping database credentials out of prompts.
- 04
Check a known value
Run a small read-only question and compare the result with the source owner's known figure.
The governed collection can answer a known question with traceable source context.
Check source health, allowed schemas or files, Space grants, and query validation before widening access.
Applied guidance
Best practices for data & business intelligence
Recommended
- Define the dataset grain, metric meaning, and data owner before asking for analysis.
- Compare a small result against a trusted source value before sharing a dashboard.
Avoid
- Do not widen schema access to compensate for a bad join or unclear metric definition.
How it fits together
01
Data Collections
Group approved structured sources into a reusable governed context rather than attaching arbitrary databases to every request.
02
Databases
Connect supported operational databases and warehouses with scoped credentials and explicit ownership.
03
Cloud data
Cloud sources remain subject to their provider permissions, network access, and organization policy.
04
MCP data sources
MCP can expose structured resources and tools while OutbornAI still applies its own source and authorization boundaries.
05
Schema indexing
Index schema metadata so the analyst can find relevant tables and columns without sending an entire catalog to the model for every request.
06
AI readiness
Clear names, stable grains, trustworthy joins, fresh data, and documented definitions improve analytical reliability.
07
Semantic context
Add business meaning around fields, metrics, entities, and relationships so analysis can use company terminology consistently.
08
Business glossary
Define important business terms and metrics separately from raw schema names.
09
Governance
Keep source ownership, access, auditability, and sensitive-data handling explicit throughout analysis.
10
Dashboards
Save reusable visual analysis into governed dashboards rather than treating every chart as a disposable chat output.
11
Reports
Reports preserve analysis and evidence for repeatable review and communication.
12
AI Analyst
AI Analyst combines governed data access, schema discovery, analysis, and visualization while retaining source context.
Related documentation