AlloyMDM
Early access

Golden records, built where your data already lives.

AlloyMDM matches the same customer, account or product across Salesforce, your marketing platform, your billing system and more, then writes one trusted record per real-world entity. It runs in your Databricks workspace on Unity Catalog tables. Your data never leaves.

The problem

The same company shows up as Acme Corp in your CRM, ACME Corporation Inc. in marketing and Acme in billing. Reports disagree, campaigns double up, and nobody trusts the numbers. Traditional MDM fixes it by copying your data into another platform. AlloyMDM does the work in the lakehouse you already pay for.

What you get

Matching across all sources

Point it at your gold tables (for example sfdc_account, sfmc_account, eb_account). It matches across every source at once, not pair by pair.

Golden records and a crosswalk

One row per real entity, plus a crosswalk from every source row to its stable mdm_id. Ids stay stable between runs.

A review queue, not guesses

Clear matches merge on their own. Uncertain pairs go to a review queue instead of being merged silently.

Explainable results

A lineage table records which source supplied each golden value and why it won.

Rules you control

Match thresholds and survivorship rules are configuration your team edits. Preview what a rule change would do before it takes effect.

Governed by Unity Catalog

Output is ordinary Unity Catalog tables. Your existing permissions apply. AlloyMDM does not invent its own roles.

Start small, grow into the app

Begin with the Job, a step you add to any Databricks job. Add the App, a steward workbench, later. Both use the same tables, so nothing needs migrating.

Compare the Job and the App →

Where we are, honestly. The matching engine and the Job are built and tested on local files. They have not yet been run on a customer Databricks workspace; that is what early access is for. The App is designed but not built yet. We are looking for a few teams to try the Job on real data.