savanna/ Build with AI
Last Updated: October 20, 2018

Accelerating graph development using Savanna's AI-assisted schema generation.

Build with AI

Savanna's AI Build feature allows you to transform raw data into an actionable graph in minutes. It automates the most time-consuming parts of graph modeling.

1. How it Works

The AI Build engine analyzes your sample data files to understand relationships and entity types.

  1. Upload: Provide CSV/TSV files with headers (Max 1GB).
  2. Analyze: AI drafts a recommended Graph Schema.
  3. Map: AI automatically maps CSV columns to vertex/edge attributes.
  4. Deploy: The system builds the schema, loads data, and installs starter queries.

2. Prerequisites & Limits

  • Data Format: CSV/TSV with clear headers.
  • Column Limit: Maximum 120 columns across all uploaded files.
  • Size: 1GB total limit for the AI-assisted flow.

3. The "Manual Edit" Option

While the AI provides a "best guess" schema, you retain full control.

  • You can review the AI-drafted schema before deployment.
  • Switch to Manual Edit mode to fine-tune primary IDs, data types, or relationship cardinality.

4. Next Steps after AI Build

Once the AI finishes the build (typically 3–5 minutes):

  • Explore: Use the Explore Graph tool to visualize the newly loaded data.
  • Query: Open the GSQL Editor to run the pre-installed analytical queries.
  • Refine: Use the Schema Designer to add more types or attributes manually.

[!TIP] AI Build is perfect for rapid prototyping and "Day 0" development. For complex production schemas, use the drafted schema as a foundation and then apply manual optimizations.