Build A HEART Eval Rubric
Apply the HEART framework to define balanced metrics for a voice AI agent.
Create a balanced evaluation rubric for an order-status voice agent whose current dashboard only shows containment. HEART for voice agents: Happiness, Engagement, Adoption, Retention, Task success. The common trap is treating containment as success. A contained call can still be confused, unresolved, or repeated later. Step 1 Name the goal: callers should get trustworthy order status without needing a repeat contact. A goal frames metrics around caller progress, not automation pride. Step 2 Map HEART signals: Happiness = complaint or rating; Engagement = repair loops; Adoption = eligible callers who continue; Retention = seven-day repeat contact; Task success = correct…
Sign up free — one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in