Fill in the blank: The three tiers of AEO metrics, ranked from most reliable to most noisy, are: directly observable metrics, metrics that require __________, and directional indicators.

The correct answer is "interpretation." AEO (AI Engine Optimization) metrics are categorized by their reliability and potential for noise. Directly observable metrics are the most reliable because they are factual and require no inference (e.g., a direct conversion). Metrics that require interpretation are less reliable because they involve some level of judgment or inference to understand their meaning or impact (e.g., sentiment analysis, which can be subjective). Directional indicators are the least reliable and most noisy, offering only a general sense of trend without precise measurement (e.g., social media mentions without context). "Automation" is incorrect because while automation can collect metrics, it doesn't define a tier of metric reliability. "Verification" is a process applied to metrics, not a tier itself. "Calibration" is a process to adjust tools or models, not a type of metric.