Duration: 5

ABINZ Orbit is a Snowflake Native App that creates a governed semantic layer on top of your Snowflake tables.

It automatically profiles your data, discovers relationships between tables, and generates a Cortex Analyst semantic model — so users can query data in plain English and receive SQL-backed results.

The app runs entirely within your Snowflake account using the Native Apps framework. No data leaves your environment.

ABINZ Orbit operates entirely within Snowflake using:

In most environments, this setup can be completed within an hour.

Duration: 2

1. Log in to your Snowflake account at app.snowflake.com
2. Click Data Products → Marketplace in the left sidebar
3. Search for ABINZ Orbit
4. Click Get on the listing
5. Choose a warehouse (e.g., COMPUTE_WH)
6. Click Get to confirm — the app installs in under a minute

Once installed, navigate to Data Products → Apps → ABINZ Orbit to open it.

Duration: 3

ABINZ Orbit can only read tables you explicitly grant it access to. Run the following as ACCOUNTADMIN:

Example using Snowflake sample data:

Duration: 2

ABINZ Orbit has two application roles:

Grant analyst access using the built-in stored procedure:

To assign admin-level access, grant the application role directly:

Duration: 1

ABINZ Orbit uses a Snowflake Serverless Task — no warehouse needed — to process metadata in the background. This is a one-time setup:

This creates a task that runs every 60 seconds to process analysis jobs in the background. The task runs on a schedule and processes jobs only when work is available.

Verify the worker is running:

A healthy output shows the task with state = started and a recent scheduled_time timestamp. If state shows suspended, re-run the START_BACKGROUND_WORKERS call.

Duration: 2

A workspace is an isolated project that holds your analysis for a specific database and schema combination.

1. Open ABINZ Orbit — you land on the Home module
2. Click Create New Workspace
3. Name it something descriptive, e.g., My First Workspace
4. Click Create

All analysis, relationships, and AI queries in this workspace are isolated from other workspaces.

Duration: 2

1. Go to the Data Sources module (left sidebar)
2. Click Browse Databases
3. Select the database and schema you granted access to in the Grant the App Access step
4. Select the tables you want to include — you can pick individual tables or include all
5. Click Add to Workspace

Once you add tables, ABINZ Orbit automatically starts analyzing them in the background.

Duration: 5

The background worker processes your tables automatically. For each table it:

Watch progress in the Data Sources module:

Analysis typically completes in 1–3 minutes per table, depending on table size and warehouse performance.

Duration: 5

Once at least two tables are analyzed, go to the Relationships module and click Discover Relationships.

ABINZ Orbit's NAME_MATCH heuristic:

This approach enables relationship discovery even when foreign keys are not explicitly defined.

High-confidence relationships are automatically accepted (default threshold: 0.7, configurable in Settings). Lower the threshold (e.g., 0.5) to surface more candidates for manual review; raise it (e.g., 0.85) to accept only the highest-certainty matches automatically.

You can:

Accepted relationships feed directly into the semantic model generation step.

Duration: 3

Go to the Model Preview module and click Generate Semantic Model.

ABINZ Orbit builds a Cortex Analyst-compatible YAML model from your tables and confirmed relationships. The model defines:

You can review and edit the YAML directly in the module before saving. The model is stored in your Snowflake account — Cortex Analyst uses it to translate plain English into SQL.

Duration: 3

Go to the AI Assistant module and type a question in the chat box. For example:

ABINZ Orbit uses Snowflake Cortex to interpret your question using the semantic model:

1. Generates a SQL query based on your intent
2. Executes it inside your Snowflake account
3. Returns results in a table and natural language summary

If the generated SQL fails (e.g., due to a schema change), the Self-Healing SQL engine attempts to detect the error, adjust the query, and retry execution using updated schema context — without any action from you.

Duration: 3

Beyond querying data, Orbit also lets you track whether insights translate into real business outcomes.

The KPI Studio module lets you define a KPI, set a target, and measure actual lift over time.

1. Go to KPI Studio → Outcome Contracts
2. Click New Contract
3. Define your KPI (e.g., "Order Fulfillment Rate")
4. Set a baseline value and a target
5. Log the intervention you are testing
6. Click Save Contract

ABINZ Orbit tracks the KPI over time and shows the measured lift — the actual change vs. your target — giving you a quantifiable record of business outcomes.

Duration: 2

Go to the Metrics Library module to find pre-defined reusable metrics. You can:

Duration: 1

You have successfully:

All of this ran entirely inside your Snowflake account. No data left your environment, and no external services were required. Your existing Snowflake governance, RBAC, and encryption policies continue to apply automatically.

Install ABINZ Orbit from the Snowflake Marketplace and connect your first dataset to get started.