Employees ask, the assistant answers — grounded in their own data.
Employee chat surface answers leave balance, attendance, probation, payslip, expense, asset, ticket and policy questions
Why teams pick AI HR Assistant
- Employee chat surface answers leave balance, attendance, probation, payslip, expense, asset, ticket and policy questions
- Grounded in the employee's own data — no hallucinated numbers, no leakage across employees
- Active policies are inline context — answers cite policy code ("per COC v2") so the source is findable
0
hallucinated numbers — answers are pulled from facts, not invented
What makes PeopleVizio's AI HR Assistant different.
Grounded by construction
Every chat call ships the employee's leave balances, last 30 days of attendance, active policies, latest payslip, open expenses, held assets and open tickets as context. The model has nothing else to draw from.
Refuses what it doesn't know
When asked something not in the context, the assistant replies "I don't have that information. Please raise a ticket or check with HR." — never invents.
Salary stays private
System prompt explicitly forbids the assistant from disclosing salary or compensation figures, even when present in the context bundle.
Concrete capabilities, tied to real code.
Every claim below maps to a specific service or class in the PeopleVizio platform.
Anthropic Claude under the hood
Stateless single-turn call per question, with prompt caching on the system prompt for cost + latency.
Graceful fallback
If the AI provider is unconfigured or unreachable, the assistant returns a polite "unavailable" message — never a stack trace.
Per-tenant control
Tenants supply their own API key via env config. Quotas and audit live with the provider.
AI HR Assistant, in detail.
Ready to see it?
Ready to see AI HR Assistant live?
Twenty minutes, tailored to your size and country, with a written summary afterwards.