Privacy and responsible AI principles
PeopleLogic helps leaders improve working conditions, staffing and management practices. It is not built to monitor employees.
Team-level, never individual
No individual attrition-risk scores, no rankings of employees, and no inference of any person's intentions, loyalty, health or personality.
Small groups are protected
An administrator-configurable minimum group size (conservative default, with a hard floor) suppresses or combines groups that could identify someone.
No identifiable survey output
Respondent identifiers are not stored. Free-text comments are redacted on a best-effort basis and visible only to authorized roles.
CVs are not used to judge people
PeopleLogic does not use CVs to infer attrition risk or make individual judgments.
Evidence is labelled
Observed indicators, potential risk signals, model estimates, AI explanations and verified findings are kept distinct, with uncertainty and limitations shown.
Correlation is not cause
Findings are hypotheses with alternative explanations and data needed to test them.
No automated employment decisions
PeopleLogic never recommends firing, disciplining or penalizing individuals or managers based on model output. Leaders remain responsible for decisions.
Access controls and audit
Role-based access, row-level security on tenant data, department-manager scoping, and an audit log of important actions.
AI with guardrails
The assistant only sees aggregated, organization-scoped results, cannot run arbitrary queries, and says when evidence is insufficient.
We do not claim certification or compliance with GDPR, NDPR, ISO 27001, SOC 2 or similar frameworks. Your organization remains responsible for lawful data collection, consent and retention.