Recruiting is where HR uses AI most — and where the legal and fairness stakes are highest. Split it into low-risk helpers and decision-shaping tools.
Low-risk helpers (start here)
- Job descriptions and ads: AI drafts from a role profile; a recruiter checks accuracy and inclusive wording.
- Scheduling: an assistant finds interview slots and sends invites.
- Candidate questions: a chat assistant answers questions about the role, process and benefits, in several languages.
- Interview guides: structured questions drafted from the job's competencies.
- Onboarding: a first-week plan, answers to "where do I find…?" questions, reminders for paperwork.
These save recruiter and hiring-manager time and improve candidate experience. The human stays in charge of every decision.
Decision-shaping tools (extra control)
- Screening and ranking of applications or assessments.
- Video or game-based assessments that score candidates.
- Matching of candidates to roles.
These are "automated employment decision tools" in New York City law and high-risk AI under the EU AI Act. They need bias testing, notices and human review (Level 2).
What good looks like
- One recruiting AI owner, a list of tools and what each one does.
- A baseline before launch: time to hire, recruiter hours per hire, application completion rate, candidate satisfaction, and selection rates by group.
- A clear rule: AI can sort the admin, people make the hiring decision.
A real example
Chipotle rolled out a conversational hiring assistant, "Ava Cado", across more than 3,500 restaurants in 2024. It chats with candidates, answers questions, collects basic information, schedules interviews and sends offers to candidates selected by managers — and Chipotle said it expected time to hire for in-restaurant roles to fall by as much as 75%.