Claire had been waiting all year for this moment. Her HR operations team had just watched a demo of AI agents resolving employee cases end to end, and her VP asked the question every HR AI lead thinks about: “So where do we start?”
She had the enthusiasm and the platform. What she didn’t have was an answer. Which of the hundreds of HR services should an AI agent take on first? Which ones had enough data & volume to learn from?
Her instinct was to pick the loudest complaint, the request everyone grumbled about in the queue. Then she remembered a lesson from an old pilot: "every successful AI agent begins with one question - Is the data ready?”
So before writing a single agent spec, she opened the HR Data Readiness Dashboard. It scored her instance across 32 metrics in six areas: operational baseline, knowledge readiness, platform readiness, self-service adoption, measurement maturity, and table health. Each metric showed up green, yellow, or red, with a ranked list of what to fix first.
The first surprise was volume. Most of her case load sat in a handful of services, and those services had the cleanest categorization. The long tail was messy: sparse history, inconsistent routing, and a lot of cases logged by HR on the employee’s behalf rather than submitted by the employee. Her biggest opportunity was not the flashiest use case. It was the one with the most volume and the most consistent data.
The dashboard pointed her to three places to start:
- High-volume, repetitive requests such as Payroll Discrepancy, Beneficiaries Add/Modify and pay statement queries. These have steady volume, predictable resolution paths, and enough historical cases for an agent to learn from
- Services with clean routing. The misrouting drill-down showed exactly which services needed cleanup before an AI agent could trust their categories.
- Knowledge gaps, knowledge articles that haven’t been updated in a while, articles with formatting issues, which weaken AI search. Fixing the top handful of HR articles would improve agent answers more than rewriting the entire knowledge base
Claire’s shortlist was no longer a guess. It was ranked by volume, data quality, and fix effort, which is exactly how she could explain the ROI to her VP.
The dashboard runs as a read-only health check. It ships with the October 2026 store release of AI Agents for HR Service Delivery, so there is nothing separate to install.
Her advice to other HR AI practitioners is simple: know before you build. Run the readiness check on your instance, fix the data that block your top use cases, and then turn on the agents. Your first win will come from the data you already have
Try it yourself: Open the HR Data Readiness Dashboard from the application navigator and click Generate Snapshot. You will need the HR AI Agents Admin role.
