The HR AI tool stack

The HR AI tool landscape is large and fast-moving. Learn the categories and their risk profiles — the specific vendors and features change quarterly (a 2026 snapshot; verify before relying on any specific, and remember: naming a tool here is not an endorsement).

The categories, roughly by risk:

Higher-risk (they influence decisions about people — Module 3's legal scrutiny applies):

• ATS with embedded AI — applicant tracking systems (Workday, Greenhouse, Ashby, Lever, SmartRecruiters, Workable) with AI matching and ranking. The ranking/matching is where legal risk concentrates.
• Sourcing / matching tools — find and rank passive candidates (Eightfold, hireEZ, Gem).
• Resume screening — parse, score, and rank applicants (a top AI use — and a top legal-risk use).
• AI interviewers / assessments — async video interviews, scoring, skills tests (HireVue, Willo). Especially scrutinized — some products dropped facial-analysis features after bias criticism.

Lower-risk (admin and support, though still with data/privacy considerations):

• Candidate chatbots — conversational screening and scheduling (Paradox's "Olivia") — strong in high-volume hiring.
• HR copilots — a conversational layer in your HRIS: draft, summarize, retrieve, and increasingly act (Microsoft Copilot for HR, vendor-native assistants).
• Onboarding / employee experience / L&D — onboarding assistants, employee Q&A, training generation.

The critical distinction — copilots vs. agents:

• Copilots suggest/draft (a human reviews and decides) — lower risk.
• Agents execute multi-step tasks autonomously — higher risk, because they act, raising the oversight and accountability stakes. As HR AI moves from copilots to agents, the "keep a human meaningfully in the loop" requirement (Module 3) becomes even more critical.

**Why think capabilities, not vendors:** the vendors, features, and product names change constantly, but the capabilities and their risk profiles are stable. A resume-screening capability carries disparate-impact legal risk whether it's called Tool A or Tool B this year. So evaluate AI HR tools by what they do and what risk that creates, not by brand — and by their compliance posture (do they support bias audits? provide adverse-impact data? — Module 4's vendor due diligence).

A key deployment principle from the start: the higher-risk categories (anything that screens, ranks, scores, or decides about candidates) require the most caution, the most legal awareness (Module 3), and the strongest safeguards (Module 4). The lower-risk categories (scheduling, drafting, Q&A) are safer starting points. Match your caution to the category's risk — and never deploy a decision-influencing AI tool without understanding its legal implications and building the safeguards.

The mindset: the HR AI stack ranges from lower-risk admin tools (chatbots, copilots, onboarding) to higher-risk decision-influencing tools (screening, ranking, AI interviewers) — and the higher-risk ones carry serious legal and ethical weight. Learn the categories and their risk profiles, evaluate tools by capability and compliance posture (not brand), start with the lower-risk uses, and treat any tool that screens or ranks candidates as requiring the full legal awareness and safeguards this course teaches. Knowing which category creates which risk is the foundation for deploying HR AI responsibly.