Anthropic CPO says 'AI FOMO' pushed companies to use the technology. Now, they want success metrics (www.businessinsider.com)

Anthropic CPO Mike Krieger told the Superhuman AI podcast that enterprise AI adoption is shifting from “AI FOMO” toward disciplined measurement: companies now want concrete success metrics or evaluations before doubling down on tools. Drawing on his product experience (he co‑founded Instagram and joined Anthropic in 2024), Krieger recommends pragmatic signals such as daily active users (DAU) as an unbiased proxy for value, and task‑specific KPIs where they exist — e.g., reduced turnaround time for legal or support tickets. He noted the harder cases where benefits are fuzzier to quantify, echoing other industry measurements like Sundar Pichai’s reported 10% engineering velocity gain, which Google framed as additional hours of engineering capacity per week from AI tooling. For the AI/ML community this signals a move to metrics‑driven deployment: model makers and platform teams must instrument products, expose measurable outcomes (time saved, error rates, throughput), and support A/B or controlled evaluations. Adoption dashboards, usage gamification and ROI calculations will guide procurement and feature prioritization, but teams should also watch for perverse incentives and privacy/monitoring concerns. In short, delivering measurable productivity or quality improvements — and clear ways to quantify them — will become a central differentiator for AI products in enterprise settings.