AutoInsights
AutoInsights analyzes your database schema using AI and automatically generates documented, priced, and publishable Spartera analytics assets — without writing a single line of SQL.
Connect a database. Review the AI's analysis. Publish to the marketplace. Start earning.
How It Works
1. Schema Analysis
AutoInsights connects to your database (via an existing Spartera connection) and reads your table structures, column names, data types, and relationships. Your actual data rows are never read during this phase.
2. AI Asset Generation
Using Gemini/Vertex AI, AutoInsights identifies meaningful analytics opportunities in your schema — revenue trends, user behavior patterns, operational KPIs — and generates SQL-based assets for each one.
3. Review & Publish
You review the generated assets, test them against your real data, adjust descriptions or pricing, and publish. Most users go from schema scan to live marketplace assets in under 30 minutes.
Getting Started
Step 1: Open AutoInsights
Dashboard → AutoInsights → New Analysis
Select an existing data connection or create a new one. AutoInsights works with all supported data sources.
Step 2: Configure Preferences
- Tables to analyze: Select specific tables, or let AutoInsights scan the full schema
- Asset types: Calculations only, visualizations only, or both
- Pricing preference: Set a default price range for generated assets
- Categories: Suggest which marketplace categories best fit your data
Step 3: Run the Free Analysis
AutoInsights scans your schema at no cost and returns a report showing:
- Number of analytics opportunities identified
- Suggested asset names and descriptions
- Estimated marketplace value per asset
- Data quality observations
Step 4: Generate Assets
Review the analysis and select which assets to generate. AutoInsights creates the full asset including SQL logic, description, tags, and suggested pricing. You can edit any field before publishing.
Step 5: Test & Publish
Each generated asset runs through the standard preview/test workflow before publishing. Once validated, publish individually or in bulk.
Tips for Better AutoInsights Results
The quality of generated assets depends heavily on how well your database is documented. A few improvements make a significant difference:
Use Descriptive Names
-- Poor: column45, tbl_x, val1
-- Better: transaction_date, customer_orders, revenue_usdAdd Column Descriptions
In BigQuery, Snowflake, and most modern warehouses you can add column-level descriptions. AutoInsights reads these and incorporates them into asset documentation.
Clean Up Test Data
Remove or isolate test/placeholder rows before running an analysis. AutoInsights uses statistical signals from your schema — dummy data skews the results.
Normalize Where Practical
Properly normalized schemas with clear foreign key relationships help AutoInsights understand entity relationships and generate more meaningful cross-table analytics.
Supported Data Sources
AutoInsights works with all Spartera-supported databases. See Supported Data Sources for the full list.
Generated Asset Types
AutoInsights creates assets across several analytics categories based on what it finds in your schema:
| Category | Example Assets Generated |
|---|---|
| Revenue & Finance | Monthly revenue trends, average order value, payment failure rate |
| Customer Behavior | Churn risk indicators, engagement scores, cohort retention |
| Operations | Fulfillment time, inventory turnover, SLA compliance rate |
| Product Analytics | Feature adoption, session duration, conversion funnels |
| Risk & Compliance | Anomaly detection, data quality scores, threshold alerts |
Pricing
AutoInsights analysis is free. Asset generation consumes credits from your account based on the number and complexity of assets generated.
| Cost | |
|---|---|
| Schema analysis | Free |
| Per asset generated | 2 credits ($1.00) |
| Bulk generation (10+) | 1.5 credits per asset |
FAQ
Does AutoInsights read my actual data?
The schema analysis phase reads only metadata (table names, column names, types). When you generate an asset, a preview execution runs against 10% of your real data to validate the SQL.
Can I edit the SQL AutoInsights generates?
Yes. Every generated asset is fully editable before and after publishing.
What if my schema has poor naming conventions?
AutoInsights will still generate assets, but quality may be lower. We recommend cleaning up naming before running an analysis for best results.
Can I run AutoInsights on multiple connections?
Yes. Run a separate analysis for each connection.
Are AutoInsights-generated assets marked differently in the marketplace?
No. Once published, they appear identically to manually created assets.
