Native Data Quality

Your data quality rules, run natively in Snowflake or Databricks.

Move your approved rules into the native data quality features the platforms now provide. No separate ETL jobs to run them, no data quality tool to buy, and your data steward gets an email when a rule fails.

Request a discovery call See how it works
The problem

Defining the rules is one thing. Running them natively in Snowflake or Databricks is another.

Your business approved its data quality rules. They still run in separate ETL jobs, or a separate tool, that read the data, test it and keep the results somewhere else. Snowflake and Databricks now have native data quality features that run these checks themselves. The features are recent, so many teams still build and maintain the separate jobs.

TodayAfter
Separate ETL jobs to build, schedule and maintain for your rules
Every rule runs natively in the platform. No separate ETL jobs
Results kept outside the platform you govern from
Results stay in Snowflake or Databricks, next to the data
Often a separate data quality tool to pay for
No data quality tool to buy. You pay the platform for what the checks use
What you get

Every rule running natively. Every failure reported.

Every rule, running natively

Each approved rule becomes a native check on its own table, with no separate ETL job to maintain.

Failures reach the right person

When a check fails, your data steward gets an email.

Quality you can show

A report on every rule: its score, its trend, and which tables and columns the rules cover.

How it works

Six weeks from approved rules to native checks in production

Your rules go into GitHub

Each approved rule goes into your GitHub repository as one versioned copy, the same in development, test and production.

AI drafts each check

An AI coding agent turns each rule into a native check on its table, and our engineer reviews every one.

The checks survive your loads

On Snowflake, a table that is dropped and rebuilt loses its checks, so we change those loads to keep them attached.

Tested, then released

We test every check against planted errors and clean data, your stewards sign off on the results, and we watch the first week in production.

Alert, report, handover

Failures email your data steward, a Power BI report shows scores and trends, and your team adds one rule on its own before we leave.

On each platform

SnowflakeDatabricks
The native checkA data metric function with an expectation, built in or customAn expectation on a table built by a Lakeflow pipeline
When it runsEvery time the table’s data changesEvery time the pipeline updates the table
The AI agentCortex CodeGenie Code, agent mode (Public Preview)
Where results goSnowflake’s own data quality tablesThe pipeline’s event log
RequiresEnterprise Edition or higherTables built by Lakeflow pipelines (Advanced edition on classic compute)
Pricing

One fixed fee. No data quality tool to buy.

$48,000

One data domain, such as Finance or Supply Chain, of up to 100 tables · no limit on rules · one platform · six weeks

  • Your approved rules as code in your GitHub, one versioned copy
  • Native checks in development, test and production, on Snowflake or Databricks
  • Load changes so every check stays attached
  • An email to your data steward when a check fails, and a Power BI data quality report
  • Two weeks of testing and release to production, then a handover where your team adds a rule

You bring

Every rule approved by your business and already written as SQL you have tested.

Add-on

Another data domain$16,000

Up to 100 more tables, their rules built, released and added to the alert and the report.

No license fee from us, and no data quality tool to buy. The checks run natively in Snowflake Enterprise Edition and in Databricks Lakeflow pipelines (Advanced edition on classic compute). You pay the platform for the compute the checks use.

Questions

What leaders ask us first

Our data quality tool already does this.
If what you need is your approved rules checked on your own tables, Snowflake and Databricks now do that natively, with no separate tool. On Databricks classic compute, that needs the Advanced pipeline edition. You pay the platform for the compute the checks use.
Are these features ready?
On Snowflake, yes. On Databricks, the checks are ready; the AI agent we use there, Genie Code agent mode, is in Public Preview. On both platforms, our engineer reviews every check the agent writes.
Our loads rebuild the tables.
On Snowflake, a table that is dropped and rebuilt loses its checks without warning. We change those loads during the build so the checks stay attached. On Databricks the checks are part of the pipeline code and deploy with it.
We don’t have approved rules written as SQL yet.
Start with Metadata Enrichment Engine. It proposes quality rules for the columns your business depends on, each with the SQL that runs it, tested on your data, and your data owner approves every one. It runs on Snowflake today.
What do we own afterwards?
The rules, the checks, the report and the runbook. Your team adds a new rule on its own before we leave. We share our AI agent instructions and templates with you; they stay ours, and we don’t maintain your copy.
What do we need in place?
Every rule approved by your business and already written as SQL you have tested. On Snowflake, Enterprise Edition or higher. On Databricks, tables built by Lakeflow pipelines, on the Advanced edition if you use classic compute. The report is built in Power BI.

Run your approved rules natively

Six weeks, one fixed fee, and every approved rule in one data domain running natively, with no separate ETL jobs to maintain.

Request a discovery call