Designing Grounded Experiences

Practical UX guidance for building on AI Query's real response fields — chart, follow-ups, streaming and cost.

AI Query returns more than an answer string — a computed chart, follow-up questions, cost and a conversation id all come back on every call. The best integrations design around those fields directly instead of treating the response as one opaque string.

#Design principles

#UX patterns that work

PatternWhy
Show birthChart alongside the readingLets users (and your own QA) see the chart the answer is grounded in — real transparency, not a black box.
Surface followUps as tappable next questionsThey are chart-relevant, not generic — a cheap way to deepen engagement with zero extra latency.
Stream the answer instead of a spinnerTokens arrive as they are produced, so perceived speed is much better than waiting for the full response. See Streaming.
Respond in the user's language14 languages are supported end to end — pass language, or let the engine auto-detect it from the question.
Keep one stable conversationId per user threadCoherent, memory-aware follow-ups instead of the user re-explaining context every turn. See Multi-turn & Conversations.

#Pitfalls to avoid