While Anthropic's AI models are recognized for their strong coding capabilities, new data suggests they are less suitable for powering real-time AI voice and chat applications for customer service compared to rivals. The issue stems from a relatively long “time to first token,” which is the delay before the model begins generating a response.
According to data from customer service software provider NiCE, this latency can cause unnatural pauses in conversation, making the models ill-suited for voice agents where immediate responses are critical. This performance lag presents a significant hurdle for Anthropic in the growing, multi-billion dollar market for AI-powered customer experience software, potentially benefiting competitors.