RecordLint

Audit customer-data quality safely

RecordLint scans Zendesk end users only after an administrator explicitly requests a scan. It detects duplicate candidates and other data-quality findings with deterministic, explainable rules. No AI or customer-data model training is used.

Read-only by design

RecordLint requests read-only Zendesk access. It never merges users, updates profiles, changes organization memberships, or executes remediation. Administrators review the evidence and make any necessary correction manually in Zendesk.

Minimized storage and data controls

Raw profile values are encrypted in a tenant-isolated temporary workspace with a maximum 24-hour TTL and earlier cleanup on cancellation or disconnect. Durable history contains minimized finding metadata and numeric record IDs, not raw contact values. CSV reports are streamed only on explicit request and are not archived by RecordLint. Administrators can review data-minimized product events, run expired-data cleanup, disconnect OAuth, or delete all tenant data. Scheduled/background customer-data scans are not available.

Data access and purpose

Using Global OAuth, RecordLint reads installation/admin details and the authorized end-user fields needed for data-quality analysis. Data is used only for administrator-requested scanning, explainable findings, reporting, security, and deletion. Identifying fields and OAuth credentials are encrypted. RecordLint does not read ticket bodies, comments, or attachments; sell data; advertise with it; send Zendesk API Data to an LLM in the MVP; train AI models; or write customer data back to Zendesk.