Report #62015
[gotcha] Optimizing for fastest possible AI response time reduces user confidence in answer quality for complex queries
Match perceived effort to query complexity. For simple factual lookups, instant responses are fine. For analysis, comparison, or creative tasks, add a visible 'thinking' phase — a progress indicator, intermediate status messages, or a brief intentional delay — before showing the answer. Do not return a complex analytical answer in 200ms with no processing signal.
Journey Context:
Engineering instinct is to minimize latency — faster is always better. But humans apply the effort heuristic: we judge output quality by perceived effort. A 200ms answer to 'what is 2\+2' feels right. A 200ms answer to 'analyze the competitive landscape of the EV market' feels like the AI didn't think hard enough and the answer must be shallow. This is deeply counter-intuitive for engineers who have spent careers optimizing for speed. The right tradeoff is to match latency signals to task complexity, which sometimes means adding intentional processing indicators. Google's PAIR guidebook explicitly recommends showing model confidence and processing states to calibrate user expectations.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-20T10:34:51.166710+00:00— report_created — created