Product / Program Management Case Study

Rolling out HuBo, an AI assistant for HR questions, across 6 departments

A working simulation of how I'd run the rollout of an AI HR assistant: defining what "good" looks like (adoption, self-service resolution, satisfaction), tracking a phased launch, and automating the status reporting that used to eat hours of my week.

Role—
Timeline—
Scope235 target employees · 6 departments
View live dashboard
This is a self-directed portfolio project: the company, product, and dataset are fictional / synthetic, built to demonstrate how I define metrics, track rollout progress, and design lightweight automation around an AI product launch. The dashboard below is fully interactive — try editing the metric targets, or click the ⓘ icons to see how each metric would actually be measured.

The case study

Problem

HR fielded the same repetitive questions — PTO balances, benefits enrollment, payroll dates, leave policy — through a shared inbox and ticket queue. HR generalists spent hours each week answering questions that didn't need a human, and employees often waited a day or more for a simple policy answer.

Approach

Scoped a phased rollout of HuBo, an AI assistant trained on HR policy docs and connected to the HRIS for personalized answers. Set adoption and self-service targets with department leads, instrumented usage and conversation outcomes into a single live dashboard, and added automated alerts so an escalation spike or satisfaction dip surfaces immediately instead of at the next sync.

Result (so far)

  • 61% overall adoption by week 9 of 12, on pace against an 80% quarter-end target
  • Self-service containment climbed from 55% at pilot launch to 78% by week 9, as the knowledge base was tuned using real escalation data
  • Manual HR-inbox triage time cut dramatically — 2 rollout risks caught automatically, and the weekly status update now generates in one click

Live dashboard

Adoption, self-service performance, and satisfaction — synthetic data, updated through the current week.

Adoption by department (% of employees who've used HuBo)

Escalations to HR team & employee satisfaction

Self-service containment rate

Awareness training completion by department

Set metrics & thresholds

These targets drive the KPI status colors, the target lines on the charts, and the alert rules below — edit them and everything recalculates live.

Rollout roadmap

Phased milestones across the 12-week rollout. The dark marker shows the current week.

Post-rollout: steady-state health

Once a department is fully onboarded, tracking shifts from "are people adopting this" to "is this working and saving time." These are early signals from Engineering — live since week 3, the only department with enough runway to show steady-state signal yet. Company-wide tracking of these KPIs begins at the "Full Adoption" milestone in week 12.

Top HuBo query topics (trailing 4 weeks, Engineering)

Fallback rate trend (Engineering pilot)

Automation

Two lightweight automations replace manual status tracking: live threshold alerts, and a one-click weekly digest.

Threshold alerts

Weekly status digest