MillerKnoll AI
Architecture · interactive

Six child agents, refactored into six skills.

Copilot Studio agents can’t migrate between experiences — so HR Studio was rebuilt, not ported. Compare the classic Studio HR orchestrator with the new-experience single agent, and hover any capability to see how it was re-expressed. Stage 1 runs a real role-grading example.

HR Studio — one agent, six skills

Experience
New (skills-based)
Shape
Single agent
Capabilities
6 skills
Grading
Confidence-scored

The re-platform: a single agent on the new skills-based experience, with no child agents. The six classic capabilities are re-expressed as inline skills, tools, and knowledge on one agent. Job grading is deterministic against MillerKnoll’s Career Framework (Support, Professional, Management, Executive), returning a confidence-scored level with title alignment, a next-best-fit, and clarifying questions when the JD is ambiguous.

Grade this role from its job description
KnowledgeSourceTool

Hover or tap an agent above to see its pipeline and what it returns for this bid.

◆HR Studio— Single agent · sole entry point

One entry point for HR Business Partners. The agent classifies the request and runs the right capability — grade a role, answer a benefits question, prep a sensitive case, or pull read-only Workday data — then composes a single grounded response. On the new experience these capabilities are skills, tools, and knowledge on one agent rather than separate child agents.

Input →
HRBPPastes a job description: “Grade this role.”
Output ←
IntentRole grading
SkillJob Grader (+ JD Signals)

Why it matters · One agent, one entry point — the HRBP never has to know which of six capabilities to invoke.

★ Grading result — confidence-scored

Career Framework level + grade, returned with a confidence score
Title alignment + next-best-fit surfaced
Ambiguous JD → clarifying questions, never a guessed grade
One agent, six skills — replatformed natively on the new experience