Billtrust CTO Warns 1% AI Error Rate Can Trigger 100 Daily Problems

Companies are moving agentic AI beyond content generation into systems that can execute accounting, payment and customer-service tasks with limited intervention. Billtrust Chief Technology Officer John Landy says that shift raises the stakes because flawed outputs can flow directly into daily operations. Clean data, strong governance and human oversight are therefore essential to maintaining accuracy, trust and control as businesses scale automation across high-volume workflows.

Landy’s latest warning is that even a 1% AI error rate can produce 100 or more problems a day when applied to enterprise-scale workloads. Billtrust argues that companies should strengthen data infrastructure, governance controls and human-review processes before expanding agentic AI deployments. The message is that dependable automation rests on the quality and management of underlying data, rather than deployment speed or the capabilities of an AI model alone.