Can leaders trust the numbers coming out of Salesforce?

Two shipped rules feed the Data Health dimension, and this page will not inflate them. Core object with no duplicate rules (high): Account, Contact or Lead carrying no duplicate rules means nothing in the org’s own metadata stops duplicates at entry — and the rule says plainly what it cannot see: a rule that exists but is switched off, or prevention living in an MDM outside the org. Duplicate record groups (low): records sharing a matched-field value that may be duplicates — a count from an aggregate query, because Horus Eye never stores record data.

When both fire together, the gap and its consequence sit side by side in the queue: no rules on Lead, and the duplicate Lead groups that let in. Around the rules, Deep Read adds evidence — field population, duplicate and matching rule reads, and report lineage that shows which numbers rest on which fields.

There is no “Data Trust Score” in Horus Eye today, and none is quoted here. A score computed from two rules would be the strongest possible good-news claim from an absence of evidence — the product refuses that, and so does this page.

This evidence is also where honest Salesforce data cleansing starts. Most CRM data quality programs fail by cleansing alphabetically — every record, every field, no priority. The duplicate groups and report-lineage evidence invert that: a data cleansing effort scoped to the fields your executive dashboards actually rest on is smaller, cheaper, and visibly done. Horus Eye isn’t a salesforce data cleansing tool itself — it never touches record data — it is the targeting instrument that tells the cleansing where to aim.

This is the data-quality problem worth an executive’s attention: not empty fields, but fragmented identity. It is also why “how does data quality affect Agentforce” has a sharp answer — an AI agent grounded on fragmented identity confidently gives three answers to one question. Data readiness comes before agent readiness; the readiness page makes that concrete.

Horus’s plan names the unguarded objects and the standard rules to activate; the merges and the identity decisions belong to owners who know the customers. That work — matching strategy, survivorship rules, the MDM conversation — is Kemisoft’s data practice, working from Horus Eye’s evidence rather than a blank whiteboard.

How do I audit Salesforce data quality?

Start with controls, not content: which core objects lack duplicate rules, which have duplicate groups, which fields feed executive reports. Content-level profiling comes after — and needs record access Horus Eye deliberately doesn’t take.

How do duplicate rules work?

A matching rule defines what counts as “the same”; a duplicate rule decides what happens at entry — warn or block. Salesforce ships standard ones for Account, Contact and Lead; activating them is usually the first fix.

Why do dashboards disagree?

Usually identity, not arithmetic: the same entity split across records, or two reports resting on different fields for the same concept. Report lineage makes the second visible; duplicate groups the first.

How does data quality affect Agentforce?

An agent answers from the data it is grounded on. Fragmented identity means confidently inconsistent answers — the worst failure mode, because it looks like the agent works.