74% of firms rolled back AI customer service agents – but Sinch says that's smart governance
Rolling back AI agents isn't failure – it's proof enterprises are monitoring properly.
A new report from communications platform Sinch challenges the narrative that most enterprises are stuck in AI pilot mode. According to the data, 62% of companies already have AI customer service agents live in production. Yet three in four (74%) have rolled back or shut down at least one deployed AI agent on governance grounds. Sinch argues that this high rollback rate is actually a sign of strong monitoring and control – mature organizations are catching failures sooner, not failing more often.
“Higher rollback rates reflect better monitoring and control, not weaker performance,” said CPO Daniel Morris. He noted a growing 'guardrail tax', where engineering teams spend most of their time building safety systems instead of improving customer experience. Looking ahead, enterprises are shifting investment priorities: 76% now invest more in trust, security, and compliance than in AI development itself (63%). Nearly all respondents (98%) plan to increase AI spending in 2026, and 86% are evaluating new communications providers to handle cross-channel context.
- 74% of companies have rolled back at least one deployed AI customer service agent.
- Sinch argues rollbacks indicate better governance, not weaker AI performance.
- 98% of enterprises plan to increase AI investments in 2026, with 76% prioritizing trust and security over development.
Why It Matters
High AI rollback rates signal smarter governance – enterprises are investing in safety, not just models.