// user behavior

Source: Search Engine Journal

AI search engines like OpenAI's SearchGPT and Perplexity are funneling users who expect instant, distilled answers directly into websites built for human browsing patterns—and most landing pages are failing the 30-second completion test. The gap isn't technical; it's a structural mismatch between how AI tools summarize content and what conversion funnels require. Brands must optimize for bot-driven traffic or lose it to competitors who do. This creates a financial incentive to redesign information architecture for speed-first experiences, prioritizing clarity and task completion over engagement metrics built for scroll depth.

Source: Lewis C. Lin’s Newsletter

Public transit generates massive revenue and ridership despite being universally despised—a rare product category where usage doesn't correlate with satisfaction. The industry's obsession with incremental UX improvements (better maps, cleaner interfaces) treats transit dissatisfaction as a design problem when the actual issue is structural: unreliable service, long waits, and lack of control. Transit apps cannot solve the operational failures—unpredictable schedules, missed connections, crowding—that make commuting miserable. Interface polish cannot fix those problems.

Source: Search Engine Journal

Google's AI Overview feature is producing measurable behavioral friction—users backtrack to organic results at nearly double the rate of control groups, suggesting the summaries often fail to satisfy initial queries. This creates a paradox for content strategy: brands can no longer assume that ranking in position one means capturing intent, since users are now systematically validating or rejecting AI-generated answers by re-engaging with the full search results. The shift moves SEO from ranking optimization toward credibility markers that survive scrutiny—bylines, data sourcing, and structural clarity become competitive advantages precisely because they're what users verify when they don't trust the summary.

Source: Techtiff

Apple is consolidating travel friction—flights, accommodations, weather, schedules—into a single device. This shifts the model from consumer travel apps competing for attention to a platform that surfaces relevant information without prompting. The consequence: travel behavior and data lock into Apple's ecosystem, and travel companies face pressure to integrate directly into iOS rather than maintain independent consumer relationships. The competition has moved from airlines and hotels to operating systems that can coordinate the travel experience.

Source: WIRED Daily

RedNote has become infrastructure. Travelers use it to plan trips, discover local businesses, and navigate experiences in ways that shape consumer behavior more directly than Western platforms do, with integration into local payment and booking systems. Chinese platforms operate as vertically-integrated commercial ecosystems rather than attention arbitrage machines. RedNote's booking and discovery layer for tourism cannot be replicated by Instagram's fragmented creator-advertiser-consumer model. A lifestyle app becoming the de facto booking layer for an entire national tourism industry reflects a different optimization: Chinese platforms prioritize GDP-generating utility over engagement metrics—the inverse of Silicon Valley's model.

Source: Seth's Blog

Seth Godin argues that current AI deployment prioritizes individual productivity—ChatGPT queries, personalized recommendations, solo content generation—inverting the internet's original promise as a connective technology. The consumer opportunity instead lies in AI systems designed for group problem-solving, collective decision-making, and shared creation. Such systems would require different product architectures and business models than today's dominant platforms offer. This positions AI as infrastructure for coordinating human intent across networks rather than replacing human judgment or effort. That market is smaller than individual subscriptions but could unlock use cases the current AI wave is not built to serve.