Customer Service AI Faces Challenges as Pay-Per-Resolution Models Rise

The customer relationship management sector is undergoing a significant transformation, shifting from human-assisted support to fully AI-driven solutions. Companies like Zendesk, Intercom, and Salesforce are pioneering new pricing models where businesses pay based on successful AI resolutions, ranging from $0.99 to $2 per solved interaction. This model promises clear ROI with no upfront costs or charges for inactive time, aligning payment strictly with problem resolution. Research from Gartner supports this trend, predicting that by 2028, at least 70% of customer service journeys will begin with conversational AI interfaces.

However, this model presents a critical challenge: defining what constitutes a "successful resolution." AI systems may mark a case as resolved once an answer is provided or the customer ends the interaction, but this does not always reflect customer satisfaction. For example, an AI might correctly state a refund policy and close the case, yet the customer may feel frustrated and alienated, damaging long-term brand loyalty. This creates a conflict of interest where AI vendors are incentivized to close cases quickly, while brands aim to build lasting customer relationships.

Recent studies indicate that aggressive AI deployment can improve efficiency metrics but often reduces customer satisfaction and loyalty. Customers frequently report feeling trapped in "bot loops" without access to human agents for complex issues. Since customer service is a key brand touchpoint for building trust, relying solely on pay-per-resolution AI models without human oversight risks eroding this vital asset. Gidi Adlersberg, Head of Business Line at AudioCodes, emphasizes that while AI can appear to solve problems, brands may pay a higher price in lost customer goodwill over time.