Search blocks by description or jump to a page.
AI response feedback row with thumbs up and down, a copy and regenerate action, and a reason picker revealed on negative feedback.
The source for every Application block is included with Pro and Team. Start a plan to copy the code.
Already on a plan? Log in
Feedback Thumbs Row sits under an Acme AI message that summarizes Q2 pipeline data: Meridian Labs at $120k and 87% close probability, Orison Health at $84k, and Cascade Retail at $61k. The thumbs-down button is shown in its active pressed state, which reveals a reason picker with three chips, Inaccurate (pre-selected), Unhelpful, and Other, plus a Submit feedback button. Copy and Regenerate ghost buttons sit on the same row. A Privacy badge panel at the bottom reads: Feedback is used to improve ranking only. It is not linked to your account data and does not expose your workspace to model training.
REASON_OPTIONS is a small const array; the active chip is identified by id comparison so the aria-pressed state is consistent across all chips. Static open state, no Radix portals.
Reach for this block directly below any AI-generated response in a chat, assistant panel, or inline answer surface. Wire the thumb buttons to your feedback endpoint and swap the reason options for your taxonomy.
A natural flow around it on an Application Pro page:
Before
After
One strong use is the post-response feedback row in a sales AI assistant. Other feedback surfaces:
Tip: the privacy note is load-bearing trust copy; keep it visible on the first negative interaction, not buried in a help article.