New hire onboarding involves dozens of tasks spread across IT, HR, facilities, and the hiring manager. Dropped tasks mean a new employee shows up without a laptop, without system access, or without knowing who their buddy is. The experience sets the tone for their entire tenure.

The Problem

Onboarding checklists are maintained in spreadsheets or wiki pages and vary by role, department, location, and employment type. Keeping these checklists current and ensuring every task is assigned and completed requires manual coordination across multiple teams.

The AI Approach

An LLM can generate a customized onboarding checklist for each new hire based on their role, department, location, and start date. It assigns tasks to the right owners, sets appropriate deadlines relative to the start date, and tracks completion.

Three-Step Build

Step 1 - Template library. Compile all existing onboarding tasks with their conditions (which roles, departments, and locations each applies to) and responsible parties.

Step 2 - Personalized generation. When a new hire is confirmed, pass their details to the LLM along with the template library. The model generates a personalized checklist with specific tasks, owners, and deadlines adjusted for the start date.

Step 3 - Task distribution and tracking. Create tasks in your project management or HR system via API. Send reminders to task owners as deadlines approach. Flag overdue items to the HR coordinator.

Where It Breaks

Edge cases like international transfers, internal role changes, or contractor conversions have unique requirements that generic templates miss. Task owners may mark items complete without actually doing them. System access provisioning often has dependencies the model cannot sequence correctly.

The Production Path

Start with your most common hire profile to validate the approach. Add completion verification (e.g., confirming system access by testing login) for critical tasks. Build a dashboard showing onboarding completion rates by department.