This release introduces Workflows for chaining automated jobs on a single schedule, a new Insights report, new intelligence providers for enrichment (Serper web search, BuiltWith, and custom MCP servers), AI-powered rule and model generation, a redesigned rules designer, and a large set of improvements and fixes across Dedupe, Cleanse, Transfer, Verify, Enrich, and Automate.
New Features
| Feature | Description |
|---|
| Workflows | Chain multiple jobs into a sequence that runs on a single schedule. A new "Workflows" page under Automate lets you build workflows in a four-step wizard (Name → Steps → Schedule → Confirm) combining Analyze, Mass Transform, Mass Merge, Mass Verify, Mass Import, Mass Delete, Mass Convert, and Sync to Salesforce actions, with drag-and-drop step reordering. Steps execute strictly in order, and by default a failed step halts the workflow. Workflows can be scheduled (once, recurring, or manual only), run on demand via "Run Now", or triggered through the API. Runs appear in the Audit log browser with a "Workflow" event type, can be stopped like any job, and send an email notification on completion with per-step results and total duration. Workflow cards show the step chain, schedule, and last run at a glance, with an enable/disable toggle to pause scheduling without deleting. |
| Insights Report | A new report answering "why is our data dirty?" with metrics prioritized for RevOps, including Duplicate Creation Rate, Source Attribution (which sources and users produce the most duplicates), Recurrence Rate, Revenue Exposure (duplicates linked to open Opportunity pipeline value), DQ Score Trend, Duplicate Pattern Distribution, Model Coverage Gap, Merge Velocity, Confidence Distribution, and Field Completeness by Source, plus dedicated panels for Source Attribution, Pattern Analysis, and Recurrence Tracking. |
| New Intelligence Providers | Three new ways to bring outside intelligence into enrichment workflows. Serper web search is available as a DataGroomr-managed connection: web fetch reads a page directly when the record already has a usable link (at no extra search cost), and web search finds the relevant pages when it doesn't. BuiltWith joins as a bring-your-own-key provider returning technographics — the technology stack detected on a domain (CMS, frameworks, analytics, e-commerce, and more) plus available company metadata (5 credits per lookup on your own key). And admins can now register a custom MCP server as a provider, so Enrichment Actions can call proprietary or internal tools the same way they call ZoomInfo, Apollo, or Claude — with OAuth 2.1 or static token authentication, per-tool enable/disable controls, connection testing, live status, and org-scoped encrypted credentials. |
| AI-Powered Data Quality Model Templates | In addition to AI-based completeness, DataGroomr can now fully generate Incomplete Records, Junk Records, and Stale Records data quality models tailored to your org via automatic discovery of the relevant fields. |
| AI-Recommended Master Record Rules | Similar to field merge rules, you can now get AI-generated recommendations for your master record selection rules, built from all fields on your object. |
Improvements
| Improvement | Description |
|---|
| Redesigned Rules Designer | The rules designer (Master, Merge, Transform, and DQ rules) has been redesigned to match the enrichment editor experience, with quick start templates (including an AI-recommended rule prompt generator), a visual Blockly designer, and a side panel of fields and compatible objects. The built-in preview is also much more capable: load real records for testing without entering edit mode, edit values inline, and have your selected fields and values remembered the next time you open the rule. |
| NPSP Merger for Cross-Object Datasets | NPSP Merger can now be enabled for cross-object datasets, so NPSP-specific relationships (Household Accounts, Affiliations, Relationships, ContactPointEmails/Phones, and others) are correctly re-parented during cross-object merges. Existing single-object NPSP Merger behavior is unaffected. |
| Flexible Record Exports | When exporting records from Dedupe, Cleanse, or Transfer, the four mutually exclusive "Include fields" options have been replaced with combinable preset toggles plus a complete, searchable, sortable field table where any fields can be selected (the Salesforce record Id is always included as a locked row). Processed records can now be exported directly from the results dialog as well. |
| Before/After Field Values in Audit | When Salesforce Field History is enabled, the "No. of Updated Fields" value in the audit log is now a clickable link that opens a table listing each updated field with its Before and After values from the merge. |
| Audit Exports Include Related Records | Exporting individual restore or merge results from the audit log browser now includes the related records information in the CSV. |
| Transfer: Remap CSV Fields on Existing Datasets | You can now change the mapping of CSV columns to Salesforce fields for existing Transfer datasets. All CSV fields from the uploaded file are available on the Mapping tab, and the Fields tab is organized into three sections, similar to Dedupe cross-object datasets. |
| Automerge Notification | When a newly created Salesforce record is automatically merged, a notification is now displayed and the dedupe component includes a link to the master record, so it's clear what happened. |
| Gemini and Copilot in BYO LLM | Google Gemini and Microsoft Copilot can now be selected as Bring Your Own LLM options. |
| Adjustable Sidebar Width | You can now adjust the width of the left sidebar, and your selection is saved so the sidebar opens to the same size the next time. |
| Model Name on CSV Confirm Step | When uploading a CSV file, the confirm step now displays the actual matching model selected (with a [default] tag) instead of the generic "Use default mode" label. |
Fixes
- DG-5711 – Fixed the NPSP rollback of accounts merged through a contact merge: associated accounts are now fully recovered along with the restored contacts, rather than being recreated without their relationships.
- DG-6130 – PersonAccount merge and rollback now correctly track the PersonContact-side reparenting Salesforce performs under the hood: audit counts include Contact-side children, and undo restores them to the rolled-back PersonAccount/PersonContact pair instead of orphaning records or failing with relation errors.
- DG-6204 – AI bulk duplicate review no longer fails intermittently on malformed model responses; the output structure is now enforced so a single bad response can no longer discard an entire batch of AI recommendations.
- DG-6210 – Attempting to upgrade a subscription while logged into a Salesforce Sandbox org now shows a clear explanation directing you to log in to your Production org.
- DG-6217 – Fixed an error alignment issue in the Log browser.
- DG-6218 – Log browser no longer marks a "Relation not found on a master record" result as a success after undoing a person account merge.
- DG-6220 – Enabling Auto Convert in a Live Dedupe dataset now requires a confirmation checkbox, consistent with Auto Merge.
- DG-6221 – Fixed a "mergeResult.forEach is not a function" error when merging groups in cross-object datasets.
- DG-6222 – Canceling an unsaved job now prompts with the Unsaved Changes dialog instead of closing silently.
- DG-6225 – Remapping a field on a Local (CSV) dataset's Fields tab now fully applies the new mapping, so subsequent duplicate/link analysis matches on the remapped field instead of silently returning 0 duplicates.
- DG-6231 – Bulk contact merges with NPSP merge enabled no longer fail with the Apex CPU time limit: work per transaction is now bounded, and an error in one batch is recorded and surfaced in the audit without aborting the whole job.
- DG-6232 – Fixed styling issues on Convert errors in the Log Browser.
- DG-6243 – AI-suggested rules render their results in paragraphs again, and the technical instructions are no longer shown in the prompt itself.
- DG-6244 – The "Not Connected" status badge on the Intelligence Providers page is now clickable, opening the provider connection flow.
- DG-6245 – Live Dedupe datasets now reflect updated records automatically, without requiring a manual analysis.
- DG-6246 – AI Recommendations are now generated for duplicate records in Cross-Object Live Dedupe datasets when the feature is enabled.
- DG-6247 – Fixed spacing in the Enrichment Action Target Mappings.
- DG-6248 – Updated the Salesforce provider setup instructions with a clearer step-by-step guide for activating the MCP server and configuring the External Client App.
- DG-6249 – Exported merge results now list related records separately for each duplicate instead of combining them into a single list.
- DG-6250 – Fixed a permission error in the Verify Campaign package when sending Email Opt Out for invalid email addresses.
- DG-6252 – The "Test with Record" modal in the Enrichment test flow now displays a scrollbar when search results exceed the visible area, so all records can be reached and selected.
- DG-6256 – The Related records column is now included when exporting rollback event results to CSV, matching the Undo event export.
- DG-6258 – The "View in Salesforce" icon in the Log browser is no longer cut off.
- DG-6259 – Restored the "Create Field" button on the Fields tab for Transfer datasets created from CSV.
- DG-6261 – The Select fields dialog in the rule designer is now wider and includes a sorting icon, matching the Create Matching Model dialog.
- DG-6262 – Rule designer sections are now collapsible.
- DG-6263 – The Export records dialog now displays the same fields as configured for the dataset (Dedupe, Cleanse, and Transfer).
- DG-6265 – Clicking "+ Add assessment" on the Dashboard now opens the appropriate dataset tab, consistent with "+ Add transformation".
- DG-6266 – Fixed an error when setting a tag on a group in a Transfer CSV dataset.
- DG-6274 – Transfer dataset mapping now prevents mapping several CSV columns to the same Salesforce field, matching the validation applied on initial CSV upload.
- DG-6278 – The AI-recommended rule template is now hidden from the Transform and DQ rule designers, where it previously displayed a warning.
- DG-6280 – Fixed an "Unsupported data provider" error when disabling, testing the connection for, or disconnecting the Serper intelligence provider.
- DG-6282 – Added the missing Workflows documentation.
- DG-6291 – Trigger creation no longer fails with "Unable to find the specified channel" for managed-package (namespaced) objects such as those installed by AppExchange packages. Standard and plain custom objects were unaffected.
- DG-6294 – Fixed a "Something went wrong" error when opening the All Records section for certain custom objects.
- DG-6298 – The DQ model filter range configured on a job is now retained after saving.
- DG-6300 – Workflow runs no longer remain stuck "In Progress" when a pre-analyze step fails due to the subscription record limit being exceeded.
- DG-6303 – Pressing Enter in a rule designer input field no longer opens the Unsaved Changes popup.
- DG-6304 – Record IDs are now shown for extra records added in the Master, Field, and DQ rule designer previews.
- DG-6305 – Fixed date/time picker issues in the rule designer preview, including clearing values and the picker icon position at different browser zoom levels.
- DG-6313 – The Exclude Duplicates option in the Field Merge Rule preview now excludes only the records that meet the rule conditions, rather than all duplicates.
- DG-6320 – Read-only fields can once again be modified in the rule designer preview for testing purposes.
- DG-6321 – The rule designer preview search field now clears after populating values from a random Salesforce record.
- DG-6323 – Clarified how to select an object field to define compatibility in the data quality rule designer when building conditions without templates.
