Chewy

AI Has a Hospitality Problem Money Can't Fix

The article argues that massive investment in AI (models and infrastructure) has not solved the human-centered aspects of service: empathy, trust, curiosity and recovery. Drawing lessons from hospitality and retail (e.g., Ritz-Carlton, Warby Parker, REI, Chewy), the author contends that LLMs and chat interfaces often deliver confident but impersonal answers, over-refuse harmless requests, and lack mechanisms for earning trust. The piece cites academic work on overconfidence and over-refusal in models and suggests the competitive edge for AI will shift from raw model intelligence to experience design that treats users as individuals and prioritizes trust and service recovery.

Five AI Blind Spots That Hurt Conversions

The article argues that generative AI improves writing speed and quality but misses core drivers of human decision-making. It lists five common AI mistakes—focusing on understanding not decisions, reducing words not cognitive load, offering choices instead of guidance, valuing persuasion over genuine trust, and optimizing individual assets rather than long-term memory—and contends that behavioral science, not better prompts, remains marketing’s biggest competitive advantage. The piece uses real-world examples (Booking.com, Amazon, Apple, Chewy, Patagonia) to illustrate behavioral principles like social proof, processing fluency, choice overload, the effort heuristic, and the peak-end rule. Published on 2026-08-10 on MarTech (owned by Semrush).

Survey: Media Buyers Unready for Agentic Revolution

Kiri Masters argues that despite widespread announcements of AI agents for media buying, most media buyers and brands lack the data infrastructure and integrations required to adopt agentic workflows. A Skai survey of 332 paid media practitioners found a mean 'agentic readiness' score of 35.7/100, with 30% unprepared, 49% exploring, and only 3% having AI embedded at scale. Retail media and CPG brands score below the 'Building' tier, and very few organisations expose agent-readable data or end-to-end orchestration. The piece highlights the gap between C-suite enthusiasm and practitioner readiness and warns that agentic advantages will depend on retailers and platforms making data accessible to agents.