Skip to main content
NewsRelease Notes

August ‘25 Release: Real-Time Deduping, Smarter AI Models, and More     

By August 27, 2025No Comments

We’re excited to share the highlights of our August ’25 release, packed with features that make it easier than ever to keep your Salesforce data clean, accurate, and trustworthy. This release introduces real-time deduplication, smarter AI-driven models, and refinements across the platform to save you time and deliver clearer insights. 

Live Deduping: Stop Duplicates at the Source 

Duplicate records don’t just clutter Salesforce, they create downstream problems like skewed reports, misrouted opportunities, and wasted marketing spend. Traditional deduplication runs on demand or on a schedule, meaning duplicates can slip in and cause damage before anyone notices. 

With Dataset Live Deduping, that changes. 

Live Deduping works like a guardrail on your Salesforce org. Every time a new record is created or updated, DataGroomr checks it instantly against your chosen model. If any duplicates are detected, you can: 

  • Flag potential duplicates for manual review 
  • Automatically merge records above a confidence threshold you set 

Enabling auto-merge means your Salesforce stays clean the moment data enters – no waiting for nightly jobs or manual runs. It’s like having a 24/7 data steward – that never misses a duplicate. 

And because Live Deduping is built on top of Salesforce CDC (Change Data Capture), it’s lightweight and efficient – listening for events and acting immediately. 

live dedupe auto merge

We’re rolling it out carefully as an add-on to ensure maximum reliability. You can request access to live-dedupe from within the app or by reaching out to us. 

Machine Learning Models Trained on Your Data, Out of the Box  

Our ML Models are now automatically trained on your own Salesforce data as soon as you sign up, not only on common datasets. Once training completes, you’ll be notified and can apply the models across standard datasets with a single click. Because the models are tuned to your unique data patterns, they detect subtle duplicates that traditional rules, and even generic AI, often miss.  

Lead Conversion Record Owner 

A new dataset setting for Leads Conversion lets you choose record ownership upon conversion. 

Lead-Conversion-Record-Owner

Smarter Imports 

Ever uploaded a CSV only to realize you were missing a field? With Create Mapped Fields in CSV Upload, you can now add mapped fields directly in DataGroomr – no need to edit your source file. 

Smarter Imports

Consistent Visuals 

The Consistent Dataset Visuals update brings dedupe dataset tiles in line with dashboard visuals. You’ll now see duplicate percentages and colors at a glance, making it easier to monitor progress. 

Consistent Visuals

Smarter Matching for AI Models 

We’ve extended a new Required Field option for ML models. This setting ensures certain fields must match before records are considered duplicates – giving you more precise control over how AI-driven matching works. 

smarter matching

Advanced Transform Rules 

For power users, Regex Support in Transform Rules makes it possible to apply “contains” and “equals” logic using regex. That means more flexible and powerful data transformations at your fingertips. 

advanced rules

Other Improvements 

Alongside new features, we’ve introduced a wide range of refinements to make your daily work smoother and more intuitive: 

  • Visual polish: The Duplicates widget now matches Salesforce’s updated color palette, and entire tiles are clickable for faster navigation. 
  • Better search: You can now search fields by internal names, making it easier to find exactly what you need. 
  • Simplified setup: Trigger names auto-generate based on your selections, saving clicks and reducing errors. 
  • Clarity at every step: Duplicate searches that return no results now show clear messages, and dashboard widgets show the total number of merged duplicates without hovering
  • Verification efficiency: Duplicate values for emails, phones, and addresses won’t consume credits multiple times, and verification results are now reused instead of re-checked
  • Transparent reporting: Models are now displayed at the top of Cleanse datasets, and triggers remember their last Salesforce CDC token. 
  • Plan clarity: Advanced preprocessing features like synonyms and stop-words are now clearly marked as Pro+ features, with lock icons and upgrade prompts. 

Why It Matters 

DataGroomr continues to simplify and automate data quality in Salesforce: 

  • Proactive: Real-time deduping stops bad data before it spreads. 
  • Smarter: Pre-trained AI models work out of the box with minimal setup. 
  • Efficient: Verification and matching enhancements save both time and credits. 
  • Polished: Visual and UX improvements make your workflows smoother and clearer. 

Your Salesforce data just got a whole lot cleaner – without extra effort. 

That’s the August ’25 release! We’re excited to see how you use these updates to perfect your data. As always, we’d love your feedback – what’s working for you, and what should we improve next? 

Happy Datagrooming! 

Ben Novoselsky

Ben Novoselsky, DataGroomr CTO, is a hands-on software architect involved in the design and implementation of distributed systems, with over 19 years of experience. He is the author of multiple publications about the design of the distributed databases. Ben holds a Ph.D. in Computer Science from St. Petersburg State University.