November '26 release introduces related record counts in master and field merge rules and in the Dedupe review dialog, extends similarity in matching models to dates, numbers and locations, a redesigned Matching Model editor, and retry from the Log Browser. It also brings faster mass merge and mass enrichment, incremental Live Dedupe updates, a rebuilt Smart Dashboard, and faster, more capable Oscar responses.
New Features
| Key | Description |
|---|---|
| DG-6578 DG-6580 DG-6581 DG-6582 DG-6585 | Related Record Counts in Master and Field Merge Rules Master Record Rules and Field Value Rules can now use the number of records related to each duplicate, such as Opportunities, Cases, Contacts, Tasks, Events and custom child objects. In the master rule editor, choose rules like "keep the Account with the most Opportunities" or "keep the record with the most related records overall" directly from the existing Sort group by, then sort by and prioritize dropdowns, where counts appear as # of Opportunities, # of all related records and so on. For more advanced logic, the Records category adds number of related [relationship] and number of all related records blocks, a new Math category adds arithmetic and round blocks, and new Sort group by value and then sort by value blocks sort by any calculated number. In the field merge rule editor, the same blocks let you take values from the record with the most related records, for example "fill empty fields from the record with the most Cases." Counts are read live from Salesforce when the rule runs and apply consistently in individual merge, mass merge, Live Dedupe auto-merge, export with the MasterId column and the duplicate-check API. Only direct lookups are counted; deleted and archived records are excluded. Counts are never written to Salesforce, and rules that don't use counts make no extra Salesforce calls. |
| DG-5326 | Related Records in the Dedupe Review Dialog A new Show Related Records switch in the review dialog's display settings adds a Related records section below the field rows, showing # of Opportunities, # of Cases, # of Tasks and other related lists for each record, plus # of all related records. The highest value in each row is shown in bold, so you can see which record carries the most history before you merge. The switch is off by default, is remembered between sessions, and can be turned on for an open group without closing the dialog. Available for Salesforce datasets in Dedupe. |
| DG-6527 | Similarity for Matching Dates, Numbers and Locations Every Similar comparison in a Classic matching model now has a setting that controls how close two values must be to count as a match. For text fields, set the minimum similarity (65% by default). Similar is also now available for date, date/time, number, currency, percent and geolocation fields, which previously only supported Exact: set a Tolerance (±) in a unit that fits the field, such as a plus-or-minus range for numbers, days for dates, minutes, hours or days for date/time, and a radius for geolocation fields. Requirements like "Opportunities with the same Name and Contact created within 5 minutes of each other" can now be expressed in a single model, with no transform rules or extra Salesforce fields. Changing a tolerance on a Live Dedupe dataset marks its index for rebuild. |
| DG-6457 | Retry Failed Operations from the Log Browser When an operation fails, you can now re-run it directly from the Log Browser. Supported operations include Import and Mass Import (Local and Sync), Merge and Mass Merge, Convert and Mass Convert, Update and Mass Update, Transform (Cleanse and Local), and Delete and Mass Delete. |
| DG-6549 | Stale and Changed-Value Indicators for Verified Records Cleanse datasets now show a stale badge (clock icon) on records verified more than six months ago, and a changed badge (≠) on records whose values in Salesforce have changed since they were verified, with the date the change was detected shown in the tooltip. Cleanse analysis now checks verified values for changes, and the grid always shows the stored verification status, so badges, filters and suggestions stay consistent. |
| DG-6586 | "Most Related Records" Quick Start Template and AI Rule Generation A new Most Related Records Quick Start master rule template keeps the duplicate with the most related records and breaks ties by most recent update. It works for every object. Describe your rule now understands phrases such as "the account with the most opportunities" or "most activities," and personalized AI recommendations can suggest a count-based master rule when your org's data makes it useful, naming the related lists used. |
Improvements
| Key | Description |
|---|---|
| DG-6653 | Redesigned Matching Model Editor The Edit Matching Model dialog for Classic models has a new layout. The field list shows each field's type, API name, comparison and weight, with tags highlighting any non-default settings such as "Min 80% similar" or "Tolerance (±) 2 hours." Clicking a field opens a settings panel that slides in from the right, grouping Match rules (comparison, minimum similarity or tolerance, Must match, blank handling, Compare across fields) and Before comparing options (Basic clean, Transform, Synonyms, Ignore words, First N characters), each with a help tooltip. Weights must now total exactly 100% to save; a distribution bar shows each field's share, and a Balance to 100% button adjusts the remaining fields proportionally. Add field opens the familiar field selection step inside the same dialog, keeping existing settings for fields you keep. The Model fields view in Dataset Settings and Datasets Comparison uses the same layout. |
| DG-6574 | Incremental Live Dedupe Updates in the Duplicates List When Live Dedupe recalculates duplicate groups, the Duplicates list now updates only the groups that were added, changed or removed instead of reloading the entire list. Scroll position, filters, sorting and the open review are preserved, and the Undo option now remains available after a merge on Live Dedupe datasets. |
| DG-6655 | Faster Mass Merge Mass merge now runs multiple merge batches in parallel and adjusts the number of parallel batches automatically, backing off when Salesforce reports record locks, governor limits or concurrent request limits. Undo and rollback information is now prepared for many groups at once instead of one group at a time. In our testing, large jobs completed roughly two to three times faster, with undo and rollback unchanged. NPSP orgs continue to merge one batch at a time. |
| DG-6632 | Faster Mass Enrichment Mass enrichment now starts the next record as soon as one finishes, instead of waiting for a whole batch to complete, and prompts that use only web search or AI can process more records at once. Providers with strict rate limits, such as ZoomInfo, keep their limits. Large web-search enrichment runs that previously took several hours now complete in a fraction of the time. |
| DG-6607 | Rebuilt Smart Dashboard Engine The Smart Dashboard now calculates its totals on a separate analytics engine, so it loads faster on large orgs and no longer slows down or interferes with merges, imports, exports and verifications. Dashboard figures update within about 15 minutes after an operation completes. Insights history, including Source Attribution, Pattern Analysis and Recurrence, is preserved. |
| DG-6538 | More Complete AI Matching Model Recommendations When AI recommends fields for a new matching model, it now takes your existing dictionaries and field grouping into account, returning a complete, ready-to-use model setup. |
| DG-6638 | Faster, More Reliable Oscar Responses Oscar, the DataGroomr AI Agent, now answers faster, completes multi-step questions more consistently and uses DataGroomr tools more reliably. |
| DG-6633 | Oscar Capability Improvements Oscar can now read and summarize your saved rules, add Wait steps when creating workflows, and build transform rules that extract a value from one field into another. When more than one dataset matches a question, Oscar lists them and asks which one you mean, and root-cause questions about duplicates now consider the dataset's matching model. Object-level field statistics are clearly labeled as such rather than presented as a dataset's figures. |
| DG-6644 | Role-Based Access in Oscar Oscar now follows the same role permissions as the DataGroomr app for reads and actions, including module access, dataset configuration, mass actions, rollback and undo. Confirmation cards never show data from datasets you can't open, and when an action isn't permitted, Oscar tells you which permission is missing. |
| DG-6443 | Clear Impact of Subscription Changes When you change your subscription, DataGroomr now shows which features, users, datasets, models, triggers and workflows will be affected before you confirm. |
| DG-6600 | Shared Credits Across Connected Orgs Connected Production and Sandbox orgs now share the same credits pool, so you can test verification runs in a Sandbox using your Production credits. |
Fixes
Merge
- DG-6637 – Fixed mass merges with Rollback enabled running out of memory and stopping at 0% when the selection included very large duplicate groups.
- DG-6641 – Fixed the Audit Log showing incomplete Groups, Records and Estimated records totals when a mass merge job was automatically retried. Totals and merge history now cover the whole job.
- DG-6634 – Fixed custom object merges, such as Campaigns, failing with a "duplicate value found" error when a unique field was filled from a duplicate record.
- DG-6635 – Fixed merges failing with "entity is deleted" when a trigger deleted and re-created a related record during the merge, as Nonprofit Cloud does with Donor Gift Summary.
Export and Transfer
- DG-6568 – Fixed exports and data loads failing entirely when the connected Salesforce user did not have access to user Profile information.
- DG-6658 – Fixed Transfer datasets ignoring Created Date and Last Modified Date values mapped from the uploaded CSV and using the current date and time instead.
Matching and Enrichment
- DG-6639 – Fixed grouped fields being skipped during manual training of ML matching models.
- DG-6654 – Fixed enrichment Target Mapping listing lookup fields by their relationship name (for example, "Parent") instead of the lookup field (for example, "Parent Account ID"), which caused tests and enrichment runs to fail.
Oscar (AI Agent)
- DG-6642 – Clicking Stop now ends Oscar's reply right away, so sending a new message no longer returns "A reply is already in progress." Changes you have approved always run to completion.
- DG-6646 – Fixed workflows created by Oscar appearing empty when opened for review right after they were created.
- DG-6651 – Oscar now correctly recognizes duplicates found by additional matching models on a dataset.
- DG-6647 – Fixed pressing Enter selecting a highlighted suggestion instead of sending the text being edited in Oscar's input field.
- DG-6479 – Images generated by Oscar can now be opened in a full-size view.