Intelligence · Module
Data Clean Rooms
Measure overlap with a partner without exchanging raw customer data
Data Clean Rooms is a long-term planned module for privacy-preserving collaboration between companies. Parties contribute tokenised data, run only approved aggregate queries under a privacy budget, and get deletion certificates when the room closes.
Collaboration without handing over customers
Two companies often want to know how many customers they share or how a joint campaign performed, without giving each other their customer lists. Data Clean Rooms is planned for W6 and requires Data Catalog & Governance, because consent-purpose tags and policies are prerequisites.
How a room works
A clean room has parties, a cross-company agreement, a purpose and a duration. Each party contributes data through hashed or tokenised identifiers, and raw data is never visible to the other parties. Queries are limited to allowed templates such as overlap counts, aggregate conversions and reach, each with minimum aggregation thresholds. Custom query templates need multi-party approval.
Privacy budget and review
Each room and party has a privacy budget. Differential privacy noise is added to results and queries are blocked once the budget is spent. Similarity checks block query sequences that look like reconstruction attempts. Every query and result is reviewed and audited. Before data is contributed, its consent basis is checked against Data Catalog purpose tags (consent basis to be verified with counsel).
Ending cleanly
When a room terminates, contributed data is deleted and each party receives a deletion certificate. Activating aggregate audiences into Marketing with consent checks is planned later. Agents may only request pre-approved templates and never propose new ones. Privacy budget handling and multi-party agreements are still partly specified in the plan.
How it fits
Clean rooms build on the Data Catalog for purpose tags and policies, on Synthetic Data Studio for safe previews, and on Marketing for the eventual activation of aggregate audiences. Every room is recorded as a cross-company agreement in the kernel, so the audit trail sits with the rest of the tenant's records.
Features
What Data Clean Rooms includes
Rooms with parties and agreements
Cross-company agreement, purpose and duration.
Tokenised contribution
Hashed or tokenised identifiers; raw data never shared.
Allowed query templates
Overlap counts, aggregate conversions and reach with minimum thresholds.
Privacy budget
Differential privacy noise; queries blocked when the budget is spent.
Consent-basis check
Contributed datasets checked against purpose tags.
Output review and audit
Every query and result reviewed and logged.
Reconstruction checks
Similar query sequences that could rebuild records are blocked.
Termination with deletion certificates
Contributed data deleted and certified per party.
Custom templates with multi-party approval
New query shapes need every party to agree.
Workflow
Run a clean room
- 1
Parties sign a cross-company agreement with a purpose
- 2
Each party checks consent basis and contributes tokenised data
- 3
Analysts run approved templates within the privacy budget
- 4
Results are reviewed and audited
- 5
At termination, data is deleted and certificates issued
AI
AI skills and Digital Employees
Reconstruction detection
Blocks queries similar to reconstruction attempts.
Privacy budget accounting
Adds noise and tracks spend per room and party.
Frequently asked questions
When is it planned?
W6 (Years 6 to 10). It is a long-term module.
Can the other party see my customers?
No. Data is contributed as tokens and only approved aggregate queries can run.
Is this legal clearance for data sharing?
No. Consent basis is checked against your tags, but the legal basis must be verified with counsel.
See Data Clean Rooms in your workflow
Tell us about your team and the outcomes you want. We will walk you through the apps, the Digital Employees and the plan for your industry.