Security & governance
Oversight, isolation, and privacy are part of the product.
Hiring involves sensitive personal data and decisions that affect people’s careers. This page describes how HireIQ is built to keep those decisions with people, keep each company’s data separate, and respect the candidates whose information it holds.
This page describes product behavior. It is not a certification, audit report, or contractual commitment.
Principles in practice
Six areas, each enforced in code.
Human decision boundary
- Models extract, organize, and explain evidence. They never reject, shortlist, select, or hire.
- Bulk close-out takes a recruiter’s proposal and a separate approval confirming the exact count.
- Candidates already under human review are excluded from bulk close-outs.
- AI suggestions record whether a reviewer accepted, edited, dismissed, or overrode them.
Tenant isolation
- Every company is a separate tenant, and every query is scoped to the viewer’s company.
- Authorization is checked on the server for each record. IDs sent from the browser are never trusted.
- API tokens embed their company scope and cannot reach another company’s data.
- Public careers pages and job feeds expose published roles only, never candidate data.
Identity and sign-in
- SAML single sign-on and SCIM 2.0 user provisioning can be configured per organization.
- Client accounts verify their email address before they can use the client workspace.
- Sign-in attempts are rate limited.
- Candidate links are 256-bit single-purpose tokens; only a hash is stored.
Candidate privacy
- Erasure deletes stored files and embeddings, redacts free text, and anonymises identity fields.
- Erasure is resumable: a failed step is recorded and retried rather than silently skipped.
- Retention rules require a documented basis and are enforced by a daily job. No default period is assumed.
- Rediscovery invitations exclude candidates who withdrew, opted out, or requested deletion.
Responsible AI and fairness
- Candidates can opt out of AI processing, and AI-assisted analysis skips them.
- Protected characteristics are never ranking signals. The copilot refuses questions about them.
- Fairness monitoring uses voluntary self-identification and the four-fifths guideline.
- Groups under 30 people are not flagged, and cells under 10 people are hidden.
Auditability and reproducibility
- Consequential actions are written to an audit log with the actor, target, and reason.
- Stage history is preserved for every application.
- Job specifications are versioned, so past matches can be reproduced.
- Evidence records keep their source and content hash, and new information never overwrites old records.
Access by role
Each workspace shows only what the role needs.
Navigation follows these rules, and so does every server-side check. Hiding a button is never the only control.
| Workspace | Role | What they can do |
|---|---|---|
| Admin | Administrator | Runs the hiring desk: requirements, pipeline, interviewers, assessments, governance, and integrations. |
| Client | Owner | Company account, team, billing, requirements, job portals, and hire decisions. |
| Client | Member | Requirements and candidates. Hire decisions follow the company setting. |
| Client | HR Head | Collects basic requirements. No candidate data, billing, or hire decisions. |
| Client | Recruiter | Requirement analysis, job-portal posting, and the candidate database. No hire decisions or billing. |
| Interviewer | Interviewer | Only the requirements and candidates assigned to them, plus their own profile. |
Application security
Defenses at the edges.
Isolated code execution
Candidate code runs only in an isolated runner with no network access. If no runner is configured, grading stops and a person reviews the answer.
Signed partner webhooks
Inbound webhooks from job portals and email intake are verified with a shared-secret signature before they are processed.
Browser protections
Every page is served with a Content Security Policy, frame protections, and a permissions policy that disables camera, microphone, and location access.
Security or privacy question?
Write to us and include your organization and the area you are reviewing.
Bring structure to your next requirement.
Create a workspace, post your first role, and see every applicant’s evidence in one pipeline.