5.2 — AI in HR (the augmented recruiter)

Many fear that AI will recruit in place of humans. Its real role: removing the paperwork to put humans back at the center.

• Skills mapping: AI analyzes thousands of internal profiles to map roles and reveal hidden talents. It amplifies internal mobility by suggesting the right position to the right employee.
• Routine automation: repetitive tasks (questions about paid leave, sorting administrative documents) move to autonomous agents. The recruiter gains time for what the machine cannot do: the interview and empathy.

🟦 Concrete no-code case — onboarding in Airtable. An Airtable base tracks each new hire; its ’‘Field Agents’’ (small AIs embedded in the table’s cells) generate the onboarding checklist, chase missing documents and summarize the manager’s feedback. The human keeps control over what matters: welcoming and deciding.

Caution

⚠️ Business guardrail (Level 4 reminder): CV screening is a domain at high risk of discrimination. A poorly designed HR agent can reproduce biases. Hence the importance of a human in the loop (HITL: Human-in-the-Loop) on selection decisions, and of auditable traces. Here, HITL is not a comfort option — it is an ethical and legal requirement.

🟢 In plain words — in HR, AI does not replace the recruiter: it removes the paperwork (chasing a document, answering ’‘how many vacation days do I have left?’’) so they spend their time where humans really matter — the interview, the decision, the welcome.

🧩 Concrete example — a 40-person SME greets each new hire with a scattered onboarding folder (contract, badge, equipment, training). An agent connected to its database automatically generates the checklist, chases the manager if a document is missing, and summarizes the end-of-probation feedback. HR no longer runs after documents: it welcomes the person. (Illustration: figures and gains vary by case.)

⚠️ Common mistake — believing a CV-screening agent is ’‘neutral’’ because it’s a machine. An agent learns from past data: if that data carries biases (gender, age, origin), it reproduces them at scale. Without a human in the loop or traces, you automate discrimination instead of correcting it.

❓ Check your understanding — Why is HITL especially mandatory on a CV-screening agent, more than on an agent that summarizes internal notes?

See the answer

Because CV screening is a decision that affects people and is legally regulated: a bias can unfairly rule out candidates (discrimination). The human and legal risk is high. Summarizing internal notes doesn’t carry that weight: an error there is fixed without serious consequences for anyone.

✅ In short

• In HR, AI removes the paperwork to give time back to humans (interview, empathy).
• Two key uses: mapping skills + automating routines.
• CV screening = high-risk domain → HITL and auditable traces mandatory.