Use cases / People and insight
AI Analytics
Turn enterprise data into organizational intelligence.
01
Problem
Business users wait on analysts for simple questions, and ad hoc exports escape data controls.
02
AI capability
Translate questions into queries, run them under the asker permissions, explain results and flag data quality issues.
03
Data
The agent works from the context you connect, not from the open internet by default.
- Data warehouse and marts
- Semantic layer and metric definitions
- Data catalog and classifications
04
Agent
Analytics Agent
Answers data questions with the query and lineage shown.
05
Workflow
- 1. Question asked
- 2. Resolve metrics
- 3. Generate query
- 4. Run with asker permissions
- 5. Explain result
- 6. Show lineage
- 7. Save for reuse
06
Governance
Governance runs inside the agent at runtime: policy changes what it can actually do.
Can
- Run read-only queries within the asker permissions
- Explain results
- Show lineage and definitions
- Suggest follow-ups
Cannot
- Write to source systems
- Export restricted columns
- Redefine metrics
- Bypass row-level rules
Requires approval
- Access to a new dataset
- Sharing results outside the organization
Records
Agent identity, Data accessed, Model used, Output, Tools called, Policy applied, Decision, Approval, Action, Outcome.
Suggested starting autonomy: L3 Supervise
L3 suits read-only querying with monitoring; metric definitions stay human-owned.
07
Outcome
What you measure, so the agent earns more autonomy on evidence:
- Analyst request backlog
- Questions answered with visible lineage
- Query errors caught
- Restricted-data requests denied
Questions
Can it see data the user cannot?
No. Queries run with the asker permissions and row-level rules.
How do we trust the numbers?
Each answer shows the metric definition, query and source tables used.