Prepare and test your skills
Prepare and test your skills
Free sample question, answered and explained below. The Generative AI Leader practice bank has 2,191 more like it. Try 10 free: instant score, no account.
An enterprise retail organization deploys a generative AI assistant to help customer service agents summarize customer histories and suggest personalized responses. The engineering team regularly evaluates the model using technical metrics such as response latency and factual accuracy.
Which metric best represents a business-centric Key Performance Indicator (KPI) to evaluate the initiative's financial and operational impact?
Business-centric Key Performance Indicators (KPIs) quantify the direct organizational impact, economic return, and strategic value generated by deploying an artificial intelligence initiative. Unlike technical benchmarks that measure how well an AI model functions computationally, business KPIs evaluate how successfully the technology achieves organizational objectives such as cost reduction, revenue growth, customer retention, and employee productivity.
Evaluating the reduction in operational cost per ticket shifts the focus from machine performance to measurable commercial value. Even a model with high computational accuracy provides limited value if it fails to improve business efficiency, shorten handling times, or reduce operational expenditure.
Factual accuracy scores and hallucination rates are model-centric evaluation metrics. They quantify the reliability, truthfulness, and grounding of a foundation model's generated content compared to reference documentation or verified source data.
Token throughput (tokens processed per second) and GPU utilization are system-level engineering metrics. They measure how effectively the underlying cloud compute infrastructure handles generative workloads and how heavily the physical or virtual accelerator hardware is being utilized during model inference.
Time to First Token (TTFT) and response latency are model-centric, technical performance metrics. They measure computational responsiveness—specifically, the elapsed time between submitting a prompt and receiving the initial output token, as well as the total duration required to complete the model's generation process.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated