Amazon rolled out an AI feature in its Shopping app that turns text prompts into custom merch—everything from hoodies to tumblers—handled end to end by Amazon’s print-on-demand service. You type a description into Alexa, it generates a design, and Amazon prints and ships it. That move puts pressure on sites like Redbubble and Bonfire by baking AI-driven product creation directly into the world’s biggest marketplace. Meta showed off an AI assistant for its Edits video app, tapping Instagram performance data to suggest clips, audio tracks and posting times. Alongside a forthcoming desktop version, it’s adding a Beta tab for experimental tools, searchable Inspiration topics and A/B testing for reels. On the phone side, iOS 27 revamps the AirPods settings screen. Instead of a long list, you get icon-driven menus up top in Settings when earbuds connect. Navigation takes fewer taps and scrolling feels more natural. UX research by Fable found participants with cognitive disabilities spotted 1.8 times more usability issues on AI-generated sites and offered twice as many suggestions as general testers. Their feedback highlighted content clarity, button labels and visual hierarchy, plus notes on mental load that other users missed. To get reliable AI prototypes, teams need an “AI-ready” design system: well-documented components, a maintained token library and auditing tools like FigmaLint to enforce consistency. Classic design principles from Dieter Rams still apply—solve real problems, reveal limitations, stay unobtrusive—while newer thinking stresses human connection over perfection. Clay’s rebrand for semiconductor search tool Partstack illustrates this mix: a simplified logo, lowercase wordmark, clear hierarchy and a visual style that pairs technical sketches with product renders to help engineers spot differences at a glance.