Properties section

Delpha’s Duplicate Detection Properties section lets you configure AI thresholds, filtering rules, screening fields, and auto-merge logic for accurate deduplication.

Duplicate Detection – Properties Section in Delpha

The Properties section of the Duplicate Detection setup allows you to fine-tune how Delpha identifies and processes duplicate records. Each setting plays a critical role in the logic and accuracy of detection.

Duplicate Properties

Key Configuration Fields

• AI Threshold Minimum confidence score for a pair to be flagged as a potential duplicate.

• Auto Merge Threshold Score from which records will be automatically merged without user review.

• Expression The initial condition to detect potential duplicates. Think of it as a formula requiring exact matches between specific fields (e.g., Name + Website).

• Active Toggle Quickly enable or disable duplicate detection for the object.

• Filtering Rule Post-detection logic that allows you to add comments or discard certain pairs based on specific field-level differences. Example: Discard if cities differ.

• Screening Fields These are the primary fields (e.g., Name) used to determine if a pair is even worth comparing. If values differ too much, Delpha skips the pair.

• Discard Values If any key field contains a listed value (e.g., "test", "dummy"), Delpha will ignore it during duplicate scoring. Syntax: comma-separated list with no spaces or quotes — e.g., text1,text2,text3

By default, the duplicate analysis will apply on all the records of the database. However, you can limit the scope using the Do Not Compare field that as been added to your org, on the targeted objects.