Salesforce admins have enough to do already. Duplicate management shouldn’t be another job.
We’ve attended a number of Salesforce community events recently, including Portugal Dreamin in sunny Lisbon last week, where we spoke with Salesforce admins, consultants and architects about duplicate management.
What became clear very quickly was that people don’t want better duplicate management processes. They want duplicate management to stop being a job.
The reaction was remarkably consistent. The mention of complex duplicate platforms, duplicate review queues, manual merges and ongoing cleanup tasks was often met with an immediate eye roll.
Not because duplicate management isn’t important, but because it’s work that people believe should already be automated. And they’re right.
The Problem Isn’t Finding Duplicates
For years, the Salesforce ecosystem has invested heavily in identifying duplicate records.
- Matching Rules.
- Duplicate Rules.
- Duplicate Reports.
- Deduplication tools.
Finding duplicates is no longer the difficult part. The real challenge begins after a duplicate has been identified.
Someone still has to:
- Review the records.
- Decide whether they should be merged.
- Determine which values should survive.
- Execute the merge.
The duplicate was found. The work comes afterwards.
Why Duplicate Management Becomes a Burden
Salesforce administrators are responsible for automation, reporting, integrations, security, user support and platform governance.
Duplicate management is rarely the reason they came to work that day.
Yet many organisations still rely on manual review processes for duplicate records.
As duplicate volumes grow, so does the administrative effort required to manage them.
The result is a cycle of:
- Find duplicates.
- Review duplicates.
- Merge duplicates.
- Repeat.
What surprised us at the community events wasn’t that people disliked duplicates. Everyone expects duplicate records to exist in a CRM.
What frustrated people was the idea that duplicate management had become an ongoing operational task.
The reaction was rarely: “I need a better way to find duplicates.”
More often it was: “Why am I still doing this manually?”
That’s an important distinction.
From Duplicate Management to Duplicate Resolution
We believe the industry is beginning to move towards a different model.
Instead of helping users manage duplicates, technology should help eliminate duplicate management.
The objective is not to create more reports.
The objective is not to create larger review queues.
The objective is not to ask users to spend more time comparing records.
The objective is to resolve duplicates consistently while requiring as little human effort as possible.
That means:
Setup the Rules
Define how duplicate records should be resolved.
Automate Resolution
Apply decisions consistently using automation.
Review Exceptions Only
Focus human attention on the small number of cases that genuinely require judgement.
The goal is simple: Reduce the number of duplicate decisions that require human involvement.
Agentforce Changes the Equation
The arrival of Agentforce introduces an interesting new possibility.
Traditionally, duplicate resolution has been a choice between automation and human review.
Now there’s a third option. Agents.
Instead of every duplicate decision being made by a user, organisations can begin delegating routine decisions to trusted agents operating within defined instructions and guardrails.
This creates a new duplicate resolution model:
Automated Resolution
Routine duplicate records are resolved automatically.
Agent-Assisted Resolution
Agents review and resolve routine duplicate decisions.
Human Review
Users focus on exceptions and edge cases.
The objective isn’t to remove control.
The objective is to remove unnecessary work.
Measuring Success Differently
Traditionally, duplicate management tools celebrate the number of duplicates they identify.
We think a more interesting question is: How many duplicate decisions required human review?
Imagine a month that looks like this:
- 1,000 duplicates identified
- 780 resolved automatically
- 190 resolved by an agent
- 30 requiring human review
If each manual review and merge takes just 2 minutes, that’s:
- 970 decisions not handled by a human
- More than 32 hours of administrative effort avoided
That’s not a duplicate management story. That’s a work elimination story. And that’s the metric that really matters.
The Future of Salesforce Duplicate Resolution
The future isn’t more complex Salesforce duplicate management platforms.
The future isn’t more duplicate reports.
The future isn’t larger review queues.
The future isn’t asking Salesforce admins to spend more time managing duplicates.
The future is fewer human decisions.
Automated resolution.
Agent-assisted resolution.
Human-reviewed exceptions.
Salesforce has become very good at identifying duplicate records. The next challenge is reducing the effort required to resolve them. Because the real problem isn’t duplicate records. It’s the work that comes afterwards.
And that’s the work we should be trying to eliminate.