An AI team is defining success criteria for a customer support chatbot.

PMI-CPMAI emphasizes establishing acceptable performance metrics and aligning AI outcomes to business value while ensuring responsible and trustworthy practices. For chatbots, business value includes deflection/containment (how many issues are resolved without human agents), customer experience (satisfaction), and operational performance (latency). Risk measures must also be included because trustworthy AI requires governance and compliance controls (privacy/security, transparency, accountability). Therefore, metrics that combine outcomes and controls—user satisfaction, containment, correct escalation/hand-off, and privacy/compliance incident rates—are the most PMI-aligned set. Response time alone (A) misses quality and risk. Features delivered (C) and lines of code (D) are delivery activity measures, not AI value or trust measures. PMI’s approach encourages metrics that support go/no-go decisions and lifecycle monitoring, making option B the best fit.