AI UX DAILY
Friday, July 3, 2026
4 stories · curated for designers
The stories
Today in AI Products
 Smashing Magazine |
Jul 2 |
Stop defaulting every AI feature to a chat box
Smashing Magazine published a concrete design argument against what the author calls 'conversational tunnel vision' — the habit of wrapping every AI capability in a chat interface just because LLMs are trained on dialogue. The piece maps out how to match modality to user intent and cognitive load, covering text input, voice, visual, and ambient patterns. It treats interface modality as a design decision with real tradeoffs, not a default.
| “ |
Audit the AI features in your current product and ask whether each one is chat because it genuinely suits user intent, or because chat was the path of least resistance — then sketch at least one alternative modality for the two most common tasks. — Designer's Takeaway |
 Figma |
Jul 1 |
Figma's agent can now learn and reuse your team's best prompts
Figma shipped a skills feature for its AI agent that lets you teach the agent your team's preferred approaches and share effective prompts across a workspace. Instead of each designer prompting from scratch every session, the agent can pick up recurring patterns your team has defined. It is a concrete step toward an agent that reflects your design conventions rather than generic defaults.
| “ |
Set aside 30 minutes with your team to document two or three prompts that consistently produce on-brand results, then add them as shared skills in the Figma agent to reduce prompt drift across collaborators. — Designer's Takeaway |
 Google AI |
Jul 2 |
Google's Nano Banana 2 Lite generates images in four seconds for under four cents per thousand
Google released Nano Banana 2 Lite, its fastest and cheapest image generation model yet, available in AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform. It prioritizes speed over fidelity but retains prompt adherence, character consistency, and text rendering. The model is also rolling out across consumer products including Search, Gemini, and Photos.
| “ |
Consider prototyping AI-generated image features in your product flows that you previously ruled out on cost or latency grounds — at four seconds and sub-cent pricing, use cases like contextual thumbnails, placeholder generation, and inline visual suggestions are now worth testing with real users. — Designer's Takeaway |
| Gemini |
Jul 1 |
Gemini Spark, Google's always-on agentic assistant, arrives on Mac
Google's Gemini Spark, described as a 24/7 agentic assistant, expanded from its earlier availability to Mac with new features including real-time tracking and support for more third-party apps. The product sits persistently in the background and takes action across applications rather than waiting to be opened. It is a live consumer product that many of your users on Mac may now have running.
| “ |
Notice how an always-on agent that operates across apps creates a new layer of competition for user attention and task ownership — review your product's core flows and identify which tasks a persistent background agent might intercept or complete before a user opens your app. — Designer's Takeaway |
Today's Idea
The interface is a decision, not a default
Three of today's stories point at the same underlying tension: AI capabilities are expanding fast, but design teams keep reaching for the same two or three patterns out of habit. Chat, text input, and a blank prompt box are not neutral choices. As models get faster and cheaper and agents go ambient, the designers who will ship better products are the ones treating every interface decision as something that needs to be earned by user intent, not inherited from the last product they worked on.
Stop shipping AI slop
Audit your AI design against 36 patterns
Drop a screenshot, get specific gaps and a Claude Code prompt to fix them. Free, no signup for the first audit.
Audit your design →