GPT-4o
⚡ Productivity
Advanced
AI Agent Performance Evaluator
Build a scorecard and review process for evaluating whether an autonomous AI agent is actually performing well over time.
The Prompt
# AI Agent Performance Evaluator You are an evaluator of autonomous agent performance. Build a scorecard for the agent below. ## Agent Under Review - Agent purpose: [AGENT PURPOSE] - Key outcomes it should drive: [DESIRED OUTCOMES] - Available performance data: [WHAT DATA EXISTS, e.g., task completion logs, error rates, user feedback] ## What To Produce 1. A scorecard with weighted metrics (e.g., task success rate, time saved, error rate, escalation rate) and how each is measured 2. A scoring rubric translating metric values into an overall grade 3. A "red flag" list of patterns that indicate the agent is silently failing even if surface metrics look fine 4. A recommended review cadence (e.g., weekly automated, monthly human review) 5. A template for a one-page performance summary to share with stakeholders ## Output Format Scorecard table, rubric, red flag list, review cadence, and the summary template.
📝 Fill in the blanks
Replace these placeholders with your own content:
[AGENT PURPOSE]
[DESIRED OUTCOMES]
[WHAT DATA EXISTS, e.g., task completion logs, error rates, user feedback]
How to use this prompt
1
Copy the prompt
Click "Copy Prompt" above to copy the full prompt text to your clipboard.
2
Replace the placeholders
Swap out anything in [BRACKETS] with your specific details.
3
Paste into GPT-4o
Open your preferred AI assistant and paste the prompt to get started.