Reasoning
While AI agent capabilities are rapidly advancing, initial widespread deployments in complex customer service scenarios often reveal significant limitations in handling nuanced or edge-case issues, leading to lower resolution rates. Current industry reports and anecdotal evidence suggest that many early adopters are still struggling to achieve consistently high resolution rates, often falling below 50% in 2025 and early 2026 disclosures, making a sub-30% rate plausible for some platforms facing particularly challenging customer segments or product complexities.Key uncertainty
The pace of improvement in large language models and their integration into customer service workflows by major platforms in the latter half of 2026.