Deduplicate Fibery Feedback Records
Fibery can surface likely duplicates through Semantic Search. Open each suggestion, compare the underlying record and either Merge or Ignore.
Before you start: Enable Semantic Search indexing for the Feedback database and check your plan’s indexed-database allowance. Preserve source identifiers before cleaning imported records.
Worked example: Two records copied from the same support ticket should usually become one. Two different customers describing a similar obstacle are separate evidence, even if their titles match.
Step-by-step guide
Choose the evidence boundary
Compare source URL, customer and date, not just the request wording. Decide whether the records are copies or independent observations.
Open the suggestion
With the database indexed, open a Feedback entity and click the duplicate badge when Fibery finds a likely match.
Inspect both records
Open the candidate and compare conflicting fields and source material. Note which customer attribution must survive the decision.
Merge or dismiss
Use Merge only for true duplicate records. Choose Ignore for a related but independent observation so it stops appearing as that duplicate suggestion.
Check the surviving record
Reopen the retained entity and verify its text, sources and relations. Confirm that independent customer evidence was not reduced to one artificial signal.
Common mistakes
Merging every similar request
Similarity can indicate a theme instead of a copy. Use feature links to group independent requests without destroying provenance.
Sources and verification
Based on product documentation checked on 2026-11-04. This is a documentation-based guide, not a hands-on benchmark. Features and availability can change.