Duplicate jobs and duplicate record sets
Key Takeaway
A duplicate job uses a matching rule to find duplicates among records that already exist in your org. A duplicate rule works differently: it alerts a user to a possible duplicate while they create a record.
Duplicate jobs scan the records already in your org for duplicates, using a matching rule you choose. A job puts its results into duplicate record sets, which are lists of records flagged as duplicates. Users can review, report on, and merge the records in those sets.
What a duplicate job does
A duplicate job uses a standard or custom matching rule to scan for duplicates across the whole org. It scans business or person accounts, contacts, or leads. You can also run duplicate jobs on custom objects.
A duplicate rule is different. It alerts a user to a possible duplicate while they create a record. A duplicate job instead searches the records that already exist in your org, so you use a job to find duplicates among existing records.