Premium membership
Continue this concept with Premium
Premium includes all published courses, graded answers, concepts and PDF guides. See the price and access terms on the pricing page.
View Premium accessone payment · six months of site access
Advanced
Perceived Latency: Making AI Features Feel Fast Before They Are Fast
Users judge speed by experience, not by end-to-end milliseconds. This guide covers techniques that reduce perceived latency in AI features (streaming, early content, progress narration for agents, skeletons, classic results first, background execution), when each applies, and how to measure perceived speed.
CONCEPT / DESIGNING FOR UNCERTAINTY
RELATED CONCEPTS
PRACTISE THIS CONCEPT
AI UX & Human-in-the-LoopWhy do AI products stream responses word by word, and when shouldn't they?→Agentic ProductsOur agent takes 90 seconds to finish a typical task and half of users leave before it's done. How do you fix it?→AI UX & Human-in-the-LoopOur AI assistant can now take actions in users' accounts. Design how users review, understand and undo what it did.→AI UX & Human-in-the-LoopDesign the experience of delegating a whole task to an AI agent, like organising a team offsite over several days within a budget.→Agentic ProductsYour team wants to build an AI agent to onboard new suppliers. Should it be an agent, or something simpler?→Agentic ProductsWhat's the difference between a chatbot, an AI workflow and an AI agent?→