Payments Cards & Mobile· Gemma Rolfe·· 14 days agoAI score50
Visa Agentic Commerce Benchmark reveals a reliability gap in AI shopping
Visa Benchmark Exposes Reliability Gap in AI Shopping
AI brief
AI-assisted source summary · not independently verified. Content standards
Visa launched Visa Agentic Commerce Benchmark. Tests showed that AI agents still struggle to consistently make reliable purchasing decisions when shopping on behalf of consumers. The tests covered 22 AI models from 12 labs and 15 shopping scenarios, with an overall pass rate of 57.2%. The best-performing model scored 82.2%, while the weakest scored 28.9%; the proportion of cases in which all three attempts in the same scenario succeeded was only 34.2%.
Source:Payments Cards & Mobile · paymentsindustryintelligence.com