Skip to main content

​
Prerequisites

Before you begin, ensure you have:
  • Prizm deployed (cloud, on-premises, or hybrid) — see Deployment Overview
  • Admin credentials to Prizm
  • Connection credentials to at least one data source (e.g., Snowflake, BigQuery, Databricks)
  • An identity provider configured for SSO (optional for quick start)

​
Step 1: Log In and Explore the UI

Navigate to your Prizm instance URL and log in with your admin credentials. You’ll land on the Home Dashboard, which shows:
  • Asset count and coverage across connected sources
  • Recent quality scores and trend indicators
  • Active exceptions and open issues
  • AI recommendations pending review

​
Step 2: Connect Your First Data Source

1

Go to Settings → Connections

Click the Settings icon in the left navigation and select Connections.
2

Click 'Add Connection'

Select your source type from the list (e.g., Snowflake, BigQuery, Databricks).
3

Enter Connection Details

Provide your host, credentials, and authentication method. Prizm supports:
  • Username/Password
  • Key Pair Authentication
  • OAuth (for supported sources)
  • Service Principal (for cloud-native sources)
4

Test the Connection

Click Test Connection to verify Prizm can reach your source.
5

Configure Discovery Scope

Select which databases, schemas, and tables to include. You can start narrow and expand later.
6

Save and Run Discovery

Save the connection and trigger the first discovery run. Assets will appear in the catalog within minutes.

​
Step 3: Explore the Data Catalog

Once discovery is complete:
  1. Navigate to Catalog in the left menu
  2. Browse or search for your discovered assets
  3. Click on any asset to view:
    • Schema and column details
    • Auto-generated descriptions (powered by Prizm AI)
    • Lineage graph
    • Profiling statistics (if available)
    • Quality scores
Use the Converse interface (chat icon) to ask natural language questions like “Show me all tables in the sales domain with freshness issues.”

​
Step 4: Run Your First Profile

Profiling establishes quality baselines for your assets:
  1. Select an asset in the catalog
  2. Click Profile → Run Now
  3. Choose profile type: Full, Incremental, Sampling, or Filter
  4. Monitor the run in Activity → Jobs
Profile results appear on the asset detail page under the Profile tab.

​
Step 5: Configure Your First Quality Metric

1

Select an Asset

Go to the asset you just profiled.
2

Click 'Add Metric'

Navigate to the Quality tab and click Add Metric.
3

Choose Metric Type

Select from Completeness, Uniqueness, Freshness, Volume, Validity, or Custom Query.
4

Set Threshold

Configure the pass/fail threshold (e.g., minimum 99% completeness for customer_id).
5

Configure Schedule

Set when the metric should run — immediately, on a CRON schedule, or triggered by a pipeline event.
6

Save and Run

Save the metric. Prizm runs the check and produces the first Score record.

​
Step 6: Enable AI Recommendations

Let Prizm suggest additional metrics and metadata for your assets:
  1. Navigate to any asset
  2. Click Prizm AI → Recommend Metrics
  3. Review the AI-suggested quality metrics with confidence scores
  4. Approve or reject individual recommendations
  5. Approved recommendations are automatically configured and scheduled

​
Step 7: Invite Your Team

  1. Go to Settings → Users & Groups
  2. Click Invite Users and enter email addresses
  3. Assign roles (Member, Steward, Owner) scoped to specific domains or resources
  4. Users receive an invitation email to join the platform

​
What’s Next?

Connect More Sources

Add dbt, Airflow, Tableau, Power BI, and more.

Configure Lineage

Enable end-to-end lineage across your data ecosystem.

Set Up Alerts

Route quality alerts to Slack, Jira, or ServiceNow.

Explore Converse

Use AI chat to manage data assets with natural language.