Skip to content

Onboarding

Qualytics combines AI-augmented data quality with the business context your team provides. After you connect a source datastore, Profile learns from observed data and creates most AI Managed checks. Observability automation creates and maintains Volumetric and Freshness checks from recorded measurements. Data stewards set the level of AI oversight, add Authored checks for business expectations, review anomalies, and decide how quality signals should drive notifications and workflows.

This guide introduces the first-run workflow from connecting data to reviewing your first quality results.

What You Need

Have the following information ready:

  • Your Qualytics deployment URL and user account.
  • The connection details and credentials for your source datastore.
  • Any network or access requirements listed in the connector documentation.
  • The business outcomes and critical data you want to prioritize first.

See Available Datastore Connectors for connector-specific requirements.

First-Run Workflow

1. Connect to Your Data

Add a source datastore to connect Qualytics to data where it already lives. A source datastore represents the database, warehouse, or object-storage location that Qualytics reads for profiling and quality checks.

Start with a focused source that supports a meaningful business process. This makes it easier to review the initial checks and anomalies with the people who understand the data.

See Getting Started with Datastores for the datastore lifecycle and setup options.

2. Sync the Structure

Run Sync to discover and map the tables, files, and fields available through the connection. Sync builds the inventory that later Profile and Scan operations use.

Review the discovered assets and confirm that the expected data is present before continuing.

See Sync for operation settings and scheduling.

3. Profile and Set AI Oversight

Run Profile to calculate statistics and learn patterns from the selected data. Profile uses those observations to generate and maintain most AI Managed checks. Volumetric and Freshness checks instead use recorded Observability measurements.

You control how this works:

  • AI Effort controls how broadly Profile explores patterns and check types.
  • Oversight determines whether new AI Managed checks become active automatically or remain drafts for review.

For critical or regulated data, keeping new checks in Draft lets a data steward review them before they become active. Your team can also create Authored checks for business expectations that cannot be inferred from data patterns alone.

See Profile and AI Managed Checks for details.

4. Scan for Anomalies

Run Scan to evaluate active checks against the selected data. Failed checks produce anomalies with information that helps your team understand what happened and which data was affected.

Treat anomalies as potential quality issues to review. Assign ownership, document the outcome, and use the resolution history to preserve the business context behind the decision.

See Getting Started with Scan and Anomalies for the next steps.

5. Act on Quality Signals

After the first Scan, use the resulting checks, quality scores, and anomalies to decide where attention is needed. You can:

  • Add or refine Authored checks for known business requirements.
  • Prioritize important datastores, containers, and fields.
  • Assign anomalies to the right owners and document resolutions.
  • Configure Flows to respond to supported operations and quality events.
  • Share quality context through integrations, the API, or MCP.

How People and AI Work Together

Responsibility What it means in Qualytics
Qualytics AI Learns from observed data and generates and maintains most AI Managed checks during Profile operations. Observability automation handles Volumetric and Freshness checks.
Data stewards and domain experts Set oversight, define business expectations, review critical checks and anomalies, document exceptions, and decide what trusted data means for each use case.
Qualytics controls Preserve checks, scores, ownership, anomaly history, and configured responses so people and systems can use the same governed quality context.

Deployment Options

Qualytics supports two deployment models:

Your deployment model does not change the core Connect, Sync, Profile, and Scan workflow.

Next Steps

  • Follow the Quick Start Guide for a screen-by-screen walkthrough.
  • Review AI Managed Checks before choosing an oversight approach.
  • Learn how Flows turn supported events into configured actions.
  • Explore AgentQ for a natural-language way to work with data quality context.