AI ran an HR role for 7 days
A real-world experiment putting an AI in an HR job for seven days found efficiency gains alongside new challenges around trust, transparency, and human judgment — a cautionary note for institutions adopting automated admin tools. The short trial underscores the governance questions campuses will face as AI handles people- and policy-sensitive tasks. (hr.economictimes.indiatimes.com)
Asanify ran a one-week live trial in which it “onboarded” an AI named Ivy as a Junior HR Executive and published the account on Apr. 1, 2026 under founder‑CEO Priyom Sarkar’s byline. (hr.economictimes.indiatimes.com)) Ivy was given a Slack account, a company email, HR admin access, set working hours and KPIs identical to a human hire to test day‑to‑day operational fit. (hr.economictimes.indiatimes.com)) Across the seven‑day run Ivy handled over 100 employee interactions — leave queries, policy questions, platform guidance and attendance inquiries — and averaged roughly a five‑minute response time versus typical benchmarks of 4–24 hours. (hr.economictimes.indiatimes.com)) The team built Ivy on a three‑layer architecture: explicit written Procedures for each workflow, a persistent Memory system to carry context across sessions, and Values & Guardrails to encode expected behavior. (hr.economictimes.indiatimes.com)) Those guardrails explicitly included company values such as “Initiate like an owner” and “Probity,” and hard escalation rules routed any sensitive or policy‑critical items straight to Priyom Sarkar for human intervention. (hr.economictimes.indiatimes.com)) Asanify framed the exercise as “eating our own cooking,” saying the process of codifying procedures for Ivy also forced the company to surface and document institutional knowledge that had previously lived in employees’ heads. (hr.economictimes.indiatimes.com))
