Data Health Scanner: find the records that need attention before trusting the report

The intended workflow is a read only assessment. Confirm the current product's permissions and behaviour before connecting a live system. Findings should lead to a separately approved correction process, especially for merges or changes that trigger automation. A data quality review should not silently become a bulk edit operation without the responsible owner's knowledge and acceptance.

Potential duplicates need human review and appropriate identifiers. Similar names or contact details can refer to different entities, while genuine duplicates may look different. Test the rules with representative records and examine false matches. The product should help prioritise investigation rather than imply that every suggested match is safe to merge automatically.

The current listing describes a widget for Zoho CRM and a web app connection route for Zoho CRM or Salesforce. Validate current availability, supported configuration and permission requirements in the walkthrough. This description does not imply a broader Salesforce implementation service or establish that every custom object and field is supported without assessment.

Bring the required module and record type rules to the walkthrough and confirm the product's current configuration options. A useful quality check reflects the process that owns the information. Avoid treating every blank field as an error or applying a rule intended for one customer category to records governed by different requirements.

No. An unusual pattern is a reason to investigate, not a conclusion about intent or misconduct. It may reflect a test record, migration, exception or legitimate business practice. Review the source and context with the owner before deciding whether correction is needed. Avoid using automated labels as an unexamined judgment about a person.

Start with the records and fields that affect a meaningful business decision, workflow or integration. Consider impact, correction ownership and the risk of changing related records. Group repeated issues to find process causes. A raw issue count alone can overemphasise easy cosmetic fixes while leaving more important identification or relationship problems unresolved.

That requires verification of the current implementation and full data flow. Review connection permissions, processing services, logs, storage and retention before using production records. Check the deployment evidence before relying on a security guarantee. The first walkthrough can use synthetic or approved redacted data while those technical arrangements are assessed.

Select one module and a set of agreed quality rules, then compare findings with a reviewed sample. Examine missed issues, incorrect flags and the effort required to resolve useful findings. Keep scanning and correction separate. The pilot should establish practical value for your process rather than promise a universal improvement in CRM accuracy.