New data reveals that AI customer service deployments are now widespread, yet many organizations face significant post-launch challenges. Three in four companies (74%) have rolled back or shut down at least one AI communication agent due to governance concerns, even as 62% operate these tools in full production.
Challenging the Pilot-Only Narrative
This trend counters the notion that most enterprises remain trapped in testing phases. Instead, the primary hurdles emerge during scaled operations, where reliability issues persist well beyond initial pilots.
Rollbacks Signal Stronger Oversight
Mature organizations exhibit higher rollback rates, which indicate robust governance practices. They identify and address failures more quickly than less advanced peers. “Higher rollback rates reflect better monitoring and control, not weaker performance,” stated Daniel Morris, Chief Product Officer at Sinch.
Morris highlighted how engineering teams devote substantial efforts to safety systems rather than enhancing customer experiences, introducing a ‘guardrail tax’ on development.
Shifting Investment Priorities
Organizations now prioritize trust, security, and compliance (76%) over core AI development (63%). Nearly all respondents (98%) plan to boost AI investments in 2026. Additionally, 86% are exploring new communications providers, while 55% develop custom infrastructure to manage cross-channel contexts.
