AI UX DAILY
Wednesday, October 7, 2026
4 stories · curated for designers
The stories
Today in AI Products
 Figma Agent |
Oct 6 |
Figma shows exactly which tasks to hand an agent and which to keep
A new Figma workflow lab shows a designer shaping a temporary-access feature while the Figma agent handles library updates, crit feedback, and a first draft of the design system spec.
| “ |
Map your next feature's tasks into 'designer-owned' and 'agent-owned' columns before you open Figma, using this workflow as the template. — Designer's Takeaway |
| Figma Motion |
Oct 5 |
Figma Motion open beta adds custom animation styles, audio, text animations, and Lottie export
Figma Motion now lets teams save and reuse custom animation styles, layer audio onto keyframes, and animate text by character, word, or line. Exports include Lottie, dotLottie, MP4, and WebM with audio.
| “ |
Audit your motion tokens this week and define at least one reusable animation style in Motion so your team stops rebuilding the same transitions from scratch. — Designer's Takeaway |
 AI Agents / Web |
Oct 6 |
Anti-bot defenses are breaking AI agents mid-task, and users bear the cost
TechCrunch reports that CAPTCHA walls, login gates, and anti-bot tools are blocking personal AI agents trying to shop or book on users' behalf, stranding tasks with no clear recovery path for the user.
| “ |
Design an explicit failure state for agent flows that hit a blocked site: show what was attempted, what failed, and give users a one-tap path to complete the task manually. — Designer's Takeaway |
| Instinct |
Oct 5 |
Instinct brings AI agents into group chats for shared planning tasks
Instinct now lets multiple people use a shared AI agent in a group chat for tasks like trip planning and carpools. Personal agents require explicit permission before sharing information or acting on someone else's behalf.
| “ |
Consider how your agent's permission model surfaces to non-account-holders in a shared context, so users can see and control what the agent shares before it acts. — Designer's Takeaway |
Today's Idea
Agents acting in public need visible permission seams
Whether an agent is updating a library, booking a trip, or acting inside a group chat, the moments where it touches something external are the moments users need control. Design those seams explicitly: show what the agent is about to do, and give users a clear way to stop it.
Stop shipping AI slop
Turn your design into Claude skills
Drop a screenshot. See which of the 38 patterns you are missing and take them away as Claude Code skills. Free, no signup for the first audit.
Was this issue worth your time?