Data and privacy: what AI sees and what it does not decide
What AI sees
| Task | What goes to the AI |
|---|---|
| Reading a resume | The resume text, the job or talent pool details, your skills list and your candidate field definitions. |
| Scoring a candidate | The resume, the job's details, your criteria and their instructions. |
| AI Job Invitation | Only the candidate's name, summary, location, experience and education level. No phone, address, ID numbers or email. |
| AI interview | The conversation, the job's questions and criteria, and your Candidate Q&A answers. |
| AI search | Only your search text. Candidates aren't sent; the search runs in your database. |
When scoring, AI is told to judge only job-relevant evidence. It must not use a candidate's name, gender, age, nationality, ethnicity, religion, marital status or appearance, or guess anything from them.
What AI does on its own
- Reads resumes and fills in candidate profiles.
- Scores applicants and talent pool members, and explains each score.
- Rates screening answers and flags knockouts.
- Labels jobs and candidates with roles, and suggests matches for jobs.
- Adds candidates to talent pools that have Allow AI Assignment on.
- Suggests screening criteria, their importance and required skills for new jobs.
What only people decide
- Which criteria count and how important they are. AI's suggestions arrive set to Must have or Nice to have; you change them as you like.
- The match levels and where they start.
- Every status change: who moves forward, is interviewed, is rejected or hired. AI never changes a candidate's status.
- Whether to trust a score. Anyone can override it with Edit next to the AI match; it then shows Manually edited.
- Whether to send anything AI wrote, such as emails and job invitations.