Two recruiters enter the same candidate weeks apart, one with a personal email on file, one with a work email. Recruiterflow's hard duplicate blocking acts as the first line of hygiene and catches exact matches on email, phone, or LinkedIn, so this pair slips through. The Duplicate Candidate Agent catches what hard blocking misses.
What it finds
The agent scans candidates in your account and groups likely duplicates using Name (First + Last) and Current Company as the starting gate. Two profiles have to match on both before they're shown at all. From there, Function and Location decide how confident the match is:
Strong Match — Name and Current Company match, and Function matches too.
Possible Match — Name and Current Company match, but Function differs.
Weak Match — Name and Current Company match, but both profiles have a Location on file and it's different. This overrides the Function check above, so a conflicting Location always pulls the pair down to Weak Match.
Job title doesn't factor into the match at any tier. Email, phone, and LinkedIn aren't part of the match either, since exact matches on those are already blocked at creation. They come back into play at merge time, as fields to combine, not signals to compare.
video
Running a scan
The agent can be run on demand every 24 hours. Either way, results land in the same review list, grouped by match tier.
Reviewing a duplicate set
For every set, Recruiterflow proposes a parent profile, the one with the most jobs associated, or if there's a tie, the profile created first. From here you can:
Merge — accept the proposed parent and combine the set.
Set Parent — override which profile the others merge into.
Dismiss — mark the set as not a real duplicate. Dismissed sets won't resurface on the next scan unless one of the profiles changes in a way that affects the match.
Higher confidence doesn't skip review. Even a Strong Match needs your explicit action, nothing merges automatically.
What happens to your data when you merge
Once you confirm a merge, Recruiterflow resolves every field and object automatically:
Multi-select fields are combined across profiles.
Single-select fields, short text, and long text are kept from the parent. If the parent's version is empty, it's backfilled from whichever duplicate was most recently updated.
Activities, and all email, phone, and social profile values, are combined, nothing is dropped.
Pipeline stage, when duplicates show conflicting stages on the same job, the stage from whichever profile moved most recently wins. A system note is added to the pipeline history explaining why, so it doesn't look like an unexplained regression later.
There's no merge-undo. Review the proposed parent and match tier before you confirm, especially for pipeline stage conflicts.
Who can use this
Access is Admin-only, unlike other agents that are open to any user.