<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Interface Report</title><link>https://interfacereport.com/</link><description>Independent publication about AI products, agents, design, Human–AI interaction, engineering, startups and SaaS.</description><language>en</language><item><title>The AI product interface is becoming a control surface, not a chat box.</title><link>https://interfacereport.com/analysis/ai-agent-control-surface/</link><guid>https://interfacereport.com/analysis/ai-agent-control-surface/</guid><pubDate>Fri, 11 Sep 2026 07:00:00 GMT</pubDate><description>As agents take action, UX shifts from prompting toward intent, permission, progress, interruption and recovery.</description></item><item><title>What makes an AI product feel trustworthy before it feels intelligent.</title><link>https://interfacereport.com/analysis/trust-before-intelligence/</link><guid>https://interfacereport.com/analysis/trust-before-intelligence/</guid><pubDate>Fri, 11 Sep 2026 07:00:00 GMT</pubDate><description>Trust is not a tone-of-voice problem. It is an interaction architecture problem.</description></item><item><title>A field guide to agent UX: what the interface should reveal, ask and remember.</title><link>https://interfacereport.com/analysis/agent-ux-patterns/</link><guid>https://interfacereport.com/analysis/agent-ux-patterns/</guid><pubDate>Fri, 11 Sep 2026 07:00:00 GMT</pubDate><description>A reusable pattern set for uncertain, asynchronous and tool-using systems.</description></item></channel></rss>