The December release brings DataGroomr to the AI tools your team already uses with Published MCP Servers and lets you describe dataset filters in plain English. It also adds two new intelligence providers, review queues in Transfer, a dedicated Integration Activity log and a long list of improvements to the AI Agent, scheduling and the Log Browser.


New Features

KeyDescription
Published MCP Servers

Publish the DataGroomr AI Agent's tools as an MCP server for your org and connect them to Claude, Agentforce or any other MCP client. Admins create servers under Setup › Integration › MCP Servers, decide who may connect, and choose exactly which tools each server exposes:

  • Reads: dataset summaries, field stats, data health metrics, matching models, duplicate groups, match explanations, source tracing, and merge, activity and workflow run history.
  • Record-level tools: verify values, check a record for duplicates before it exists in Salesforce, enrich a single record and check job status.
  • Actions: analyze a dataset, merge or undo merges, run workflows and schedules, and run mass verify, enrich and transform.
  • Configuration (optional): create datasets, matching models, rules, data quality models, enrichment actions and automations.

Every call runs as the connecting user's own Salesforce identity, so roles, dataset permissions and licenses apply exactly as in the app. Users connect with a personal token or by signing in with Salesforce, and in-product guides cover Claude Code, the Claude API, Agentforce and other clients. Write actions run once the AI client's own approval prompt is confirmed. Tools are published with a datagroomr_ prefix so they are easy to recognize next to other connectors, and a single shared connection address lets DataGroomr be added from AI providers' connector directories. Available on Trial and Enterprise plans.

AI Filter Assistant

Describe a dataset filter in plain English and DataGroomr writes the SOQL. In Dataset Configuration → Filter, SOQL Mode is renamed Advanced Mode, and the new Ask AI button turns requests such as "Contacts in the US with no email, created in the last 90 days" into a validated filter.

  • Type {{ to pick fields, including related fields.
  • Review the summary, the SOQL and any assumptions, then refine or apply.
  • Filters on related records are supported, e.g. "contacts with open opportunities" or "accounts with no cases".
  • Available on all plans.
Integration ActivityA new Audit page listing every call made by connected AI clients (MCP) and Ingestion API integrations: time, source, server or API key, user, IP address, call, status and duration. Filter by source, server or key, user, status and date, and expand a row to see details and links to the operations it started in the Log Browser. Integration records, including API job history, are kept for one year.
SMARTe Intelligence ProviderConnect your SMARTe account under Intelligence Providers and use its person, mobile, company, technographics and funding data in enrichment actions. Credits come from your SMARTe balance and are charged only when data is returned. Technographics data is returned for the technologies your SMARTe admin enables under Org Controls.
Coresignal Intelligence ProviderConnect your Coresignal account with a secure sign-in and enrich records with company, person and job data plus verified work emails. Credits come from your Coresignal balance.
Review Queues in TransferResults from Manual enrichment actions and verification suggestions on Transfer datasets go to the same review queues as Cleanse: a Pending changes bar with Enrich and Verify chips, badges on records, Apply / Skip / Ignore, edit before apply, and Apply all. Applied values on CSV rows go out with the next push, and unapplied changes are never pushed to Salesforce.


Improvements

KeyDescription
AI Agent
  • Update, pause and delete scheduled jobs, workflows, triggers, matching models, enrichment actions, datasets, rules and data quality models, with a confirmation card showing what changes and what depends on it.
  • Create cross-object Dedupe datasets, e.g. Leads in Contacts.
  • Merge a list of record IDs, the same way as selecting records in All Records, and choose the master.
  • Open an existing rule, revise it while keeping its field lists, show the changes and save after you approve.
  • Questions about where a dataset's duplicates come from return that dataset's own source breakdown.
  • Answers drawn from DataGroomr documentation include links to the articles quoted.
AI "Most Complete" Master Record TemplateThe Most Complete template in the master record rule designer uses AI to pick 3–5 key business fields for the object and builds a rule that prefers the record with those fields filled in, most important field first, falling back to the oldest record.
Time Zones on SchedulesWhen creating or editing a scheduled job or workflow, the Run at field shows its time zone, e.g. Eastern Daylight Time (UTC−04:00), with a Use my time zone link when the schedule was saved in another zone. Editing a job or workflow keeps its last run, its On/Off state and its original start date.
Log Browser: User and SourceTriggered By is split into User and Source columns. Source shows what started each operation (Job, Trigger, Workflow, API, MCP, Assistant, or the DataGroomr Duplicates, Verify or Enrich Salesforce component). Filter by user and source together, and the CSV export includes Source. Running operations show a progress bar, a live duration and a Stop button.
Log Browser Live UpdatesNew events update the Log Browser in place without reloading the whole page.
Company Preferences: Turning Off AITurning off AI in Company Preferences blocks all new AI use across the app, including the AI Agent, AI rule building, enrichment plan generation, AI duplicate review, AI name suggestions and new MCP servers. AI buttons stay visible with a message explaining how an account admin can turn AI back on. Scheduled workflows and existing configuration keep running.
Faster Mass Merge for Person AccountsMass merges on Account in Person Account orgs spend far less time collecting related records before each merge batch, so large jobs finish sooner.
Verification Suggestions in Dedupe and TransferApplying verification suggestions in Dedupe and Transfer uses the same streamlined dialog as Cleanse.
Pending Changes Bar in CleanseThe Pending changes bar sits under the toolbar and is highlighted, matching the integrity check.
Clearer Advanced SettingsAdvanced settings in dataset settings (Dedupe, Transfer, cross-object and Cleanse datasets) and in trigger settings use clearer option labels and explanations.
Unused Sandbox CleanupSandbox connections with no sign-in for 180 days are scheduled for removal. Sandbox users and the production account's administrators receive an email 30 days ahead and a reminder 7 days before; signing in to the sandbox once cancels the removal. Records in Salesforce and production DataGroomr accounts are never affected.
Consistent Close ButtonsClose (X) buttons across panels, popovers, review popups, notifications and banners share one consistent design, with dark mode and keyboard focus support.


Fixes

  • DG-6758 – Recurring scheduled jobs and workflows run on the weekday and time picked in the user's time zone, stay on time after daylight saving changes and delays, and Monthly schedules run on the same date each month. Open and save an existing schedule to apply your time zone to it.
  • DG-6673 – Live Dedupe groups records that are created or edited while a full analysis of the dataset is running.
  • DG-6750 – After a streaming interruption of up to about three days, Salesforce changes made during the interruption are replayed to Live Dedupe and triggers.
  • DG-6754 – Cancelling an analysis leaves existing duplicate results intact, and the cancel message names the user who pressed Stop.
  • DG-6788 – Mass Review continues to the next page of groups with correct numbering after every group on the current page has been merged or has failed.
  • DG-6789 – "Most common value" in field merge rules recognizes matching addresses.
  • DG-6688 – AI-generated field merge rules apply ranked preferences ("prefer A, then B, then C") in the order listed, and "equals" comparisons handle letter-case differences in values.
  • DG-6687 – Source Attribution is populated for datasets whose filters use sub-selects.
  • DG-6714 – Auto-Select after a manual field pick in Create Matching Model sets weights that total 100%.
  • DG-6665 – Rule preview counts related records correctly for records loaded from Salesforce.
  • DG-6771 – Generating a field merge rule based on related record counts (e.g. "the record with the most Cases") completes without the "Unable to count related records" error.
  • DG-6781 – Regenerating an enrichment action's plan keeps the existing field mappings.
  • DG-6821 – The multiselect field dropdown in the rule editor stays open while checking several fields.
  • DG-6790 – Signing in when the Salesforce org has exceeded its API request limit completes, or shows a clear explanation instead of a generic error.
  • DG-6648 – AI Agent fixes for answers about fields used in master record rules, transform rule generation and SOQL filter corrections.
  • DG-5560 – The Unsaved Changes popup does not appear when cancelling an unchanged Lead in Contact dataset created at first sign-in.