Billtrust CTO Says a 1% AI Error Rate Can Create 100 Problems a Day

As retailers and service providers rush to integrate AI into their payment and financing workflows, there is a growing warning about the hidden costs of automation. Billtrust CTO John Landy highlights that even a 'highly accurate' AI system with a 1% error rate can result in 100 manual corrections per day for a business processing 10,000 transactions. This is particularly critical for businesses offering consumer financing, where errors in credit applications, payment processing, or customer billing can lead to regulatory headaches and damaged customer relationships. The shift from simple chatbots to 'agentic AI'—systems that can make decisions and trigger actions autonomously—means that clean data is no longer optional. For a merchant, an AI agent that incorrectly flags a customer’s payment status or miscalculates interest due to messy data can create a snowball effect of customer service calls. Operators are encouraged to maintain 'human-in-the-loop' oversight. This ensures that while AI handles the heavy lifting of data processing, a person is still verifying high-stakes financial decisions. Before fully automating your financing workflows, ensure your underlying data is organized and that your team has a clear governance plan to catch the inevitable 1% of errors before they reach the consumer.

Source: PYMNTS

← Return to the library· Submit a correction