Intelligence · Module
BI Studio
A semantic layer and certified metrics that every team, tool and agent reads
BI Studio is the planned semantic layer for One3: models generated from app manifests, certified metric definitions, explorations and natural-language questions answered through metrics only. Dashboards, Copilot, Sheets and agents all query the same definitions.
Metrics defined once
BI Studio is planned for W3 as the semantic layer behind every number in One3. Instead of asking an analyst to model raw tables, it generates semantic models from Pack manifests and the shared schema: joins, keys, time dimensions and currency handling. When a Pack upgrades, the model regenerates with a diff for review.
Certification you can trace
Each metric has a measure, aggregation, filters, time grain, allowed dimensions, unit, currency rule, owner, version and a status: draft, in review, certified or deprecated. Certification needs Metric Owner approval with segregation of duties, and shows an impact list of the dashboards, alerts and agents that use the metric. Deprecating a metric requires dealing with every dependant first. Lineage traces each metric to source objects, read models and datasets. Only certified metrics can be used by agents at A2 and above and by the Decision Engine.
Explore and ask
Explorations let users pick metrics and dimensions, pivot, filter, chart, drill, save and share. NL to metric parses a question into metrics, filters, grain and comparison, shows the interpreted query for editing, and answers through metrics only, never raw SQL against tables. Every result shows a data-quality score and freshness, and answers are refused below the threshold with a fix list.
Analysis depth
Anomaly explanation uses contribution analysis across dimensions, labelled "where, not why", with control for multiple comparisons. Calculated metrics are validated for formulas and units, and mixed-currency aggregation needs an explicit conversion rule. Metric trees (for example revenue = leads × conversion × average order value) feed the Decision Engine and Forecast & Simulation. Paginated reports, cohort analysis, targets from Goals and clustering into segments are planned.
Security and reach
Row-level and column-level security come from app permissions and field-level security; query caches are keyed by permission scope. The metric.query tool over API and MCP returns the definition version, freshness and quality with every answer. Sheets connects to the semantic layer, data agents can be bound to a small set of models, and an import assistant for Power BI, Looker and Tableau definitions is planned for W4, without a promise of exact parity.
Features
What BI Studio includes
Generated semantic models
Built from Pack manifests and the shared schema, regenerated on upgrade with a diff.
Metric definitions and certification
Draft, in review, certified or deprecated, with owner approval, SoD and an impact list.
Dimensions and hierarchies
Fiscal calendars, India states and districts, org units, product categories, sub-accounts and legal entities.
Explorations
Pivot, filter, sort, chart, drill, save and share.
NL to metric query
Editable interpretation; answered via certified metrics only, never raw SQL to tables.
Quality score on every result
Freshness and data quality shown; answers refused below threshold with a fix list.
Metric lineage
From source objects and datasets through metrics to dashboards, alerts and agents.
Anomaly contribution analysis
"Where, not why" decomposition across dimensions with multiple-comparison control.
Calculated metrics
Ratios, differences and conditional counts with formula and unit checks.
Row and column security
Inherited from app permissions and field-level security; cache keyed by permission scope.
Metric trees
Driver decompositions used by the Decision Engine and Forecast & Simulation.
Reports
Paginated tabular reports with subtotals, parameters and scheduled delivery.
metric.query API and MCP tool
Returns definition version, freshness and quality with every result.
Data agents and Sheets connection
Read-only conversational agents bound to a few models; spreadsheets connected to the semantic layer.
Workflow
Metric lifecycle and query
- 1
A semantic model is generated from Pack manifests
- 2
A metric is drafted and submitted for review
- 3
The Metric Owner certifies it, seeing every dependant
- 4
Dashboards, Copilot, Sheets and agents query the certified metric
- 5
Results carry definition version, freshness and quality
- 6
Changes go back through review with an impact list
AI
AI skills and Digital Employees
NL to metric query
Constrained metric query with clarification and an editable interpretation.
Semantic model drafting
Generates model definitions from manifests; certification is human.
Formula assistant
Helps write calculated metrics with unit checks.
Definition conflict detection
Finds conflicting metric definitions and drafts change proposals.
Digital Employees that use this module
They act through the same governed Actions as people; risky steps wait for approval.
Frequently asked questions
Do I need BI Studio for dashboards?
No. Dashboards use the metric registry from W1. BI Studio, planned for W3, adds a full semantic layer, custom certified metrics and explorations.
Does natural language query write SQL against my tables?
No. Questions are answered through certified metrics only, and the interpreted query is shown for editing.
Can I import my Power BI or Looker measures?
An import assistant with a reconciliation report is planned for W4. Exact parity is not promised.
See BI Studio in your workflow
Tell us about your team and the outcomes you want. We will walk you through the apps, the Digital Employees and the plan for your industry.