AI AGENTS

AI Agents Face Major Backlash in Customer Service Deployments

Recent research shows 74% of firms have rolled back AI agents in customer service, highlighting significant governance and operational issues despite increasing investments.

AI Agents Face Major Backlash in Customer Service Deployments
CoinSynaptic Desk
AI AGENTS · Correspondent
· PUBLISHED MAY 18, 2026 · UPDATED 12:07 ET · 3 MIN READ

A significant shift is underway in customer service, with a new report revealing that 74% of companies have either shut down or rolled back their AI agents. This trend raises important questions about the effectiveness of AI in customer communications and the governance structures that support these technologies. The research, commissioned by Sinch and based on a survey of over 2,500 industry leaders, shows that while AI agents are moving beyond pilot phases, their deployment is fraught with challenges.

High Rollback Rates Indicate Governance Failures

Despite initial enthusiasm for AI agents, many organizations are now reconsidering their approach due to performance issues. Nearly 66% of respondents reported having AI agents operational within their organizations, and 88% plan to move these agents into full production within the next year. However, governance failures are the primary reason behind the decision to roll back AI customer communication agents. Concerns over data exposure, hallucinations—where AI provides incorrect information—and a lack of auditability have raised alarms among companies.

The report notes that 31% of organizations cited customer data exposure as a major risk, while 22% expressed concern about potential brand damage from incorrect AI outputs. "These aren’t abstract risk categories," the Sinch report stated, highlighting the real consequences of inadequate governance.

Governance Maturity Not Enough to Mitigate Risks

Interestingly, the survey found that rollback rates were even higher among organizations claiming to have mature governance protocols. A staggering 81% of these firms reported having to retract AI agents, suggesting that better monitoring may lead to quicker identification of failures rather than improved performance. Daniel Morris, CPO at Sinch, remarked, "The industry has assumed that better governance leads to better outcomes, but that's not enough. If governance was the fix, the most mature teams would roll back less, not more."

See also  AI-Driven Fraud Escalates, Threatening Digital Trust in Finance

These findings indicate that advanced organizations are not necessarily failing less; they are simply detecting failures more quickly. This could reflect a more proactive approach to governance but also points to persistent underlying issues.

Investment in Trust and Security

In response to these challenges, companies are shifting their AI investments toward building trust, security, and compliance measures rather than focusing solely on AI technology. According to the survey, 75% of organizations prioritized investments in these areas, compared to only 63% for AI technology. Additionally, 84% of AI engineering teams reported spending at least half their time on safety infrastructure, diverting resources from enhancing customer experience.

Illustrative visual for: AI Agents Face Major Backlash in Customer Service Deployments

Morris emphasized the implications of this "guardrail tax," stating, "Engineering teams are spending most of their time building and maintaining safety systems, a lot of which their communications infrastructure should be providing, instead of focusing on improving the customer experience."

Despite these setbacks, interest in AI for customer service remains strong. Chatbots and automated email responses are the most widely adopted applications, used by over 60% of respondents. Other channels, such as social media and messaging platforms, are also seeing increased adoption, with half of the surveyed organizations utilizing AI agents in these areas. The main reasons for such deployments include enhancing customer satisfaction (36%), driving revenue and conversions (24%), and reducing operational costs (16%).

As companies continue to face governance and operational challenges, their commitment to AI investment remains firm. The drive to integrate AI into customer service operations is likely to persist, though it will require a careful reassessment of governance frameworks to address identified risks. Balancing technological advancement with stable governance will be crucial as organizations navigate the complexities of AI implementation in customer service.

See also  Okta Enhances AI Agent Governance with AWS Integration

With AI agents under scrutiny, the industry finds itself at a critical juncture. The recent findings highlight the urgent need for businesses to refine their strategies, ensuring that the deployment of AI technologies aligns with operational goals and fosters customer trust. Moving forward will require a concerted effort to establish a sustainable balance between innovation and accountability in AI-driven customer service.

Quick answers

What percentage of companies have rolled back their AI agents?

74% of companies surveyed have either shut down or rolled back AI agents used in customer service.

What are the main reasons for AI agent rollbacks?

The leading factors include customer data exposure (31%), hallucinations (22%), and lack of auditability (16%).

How much time are AI engineering teams spending on safety infrastructure?

84% of AI engineering teams spend at least half their time on safety infrastructure.

What are the primary motivations for deploying AI agents?

The main motivations are improving customer satisfaction (36%), boosting revenue and conversions (24%), and reducing operational costs (16%).

CoinSynaptic Desk

AI Agents · 2,404 stories

CoinSynaptic Desk covers the intersection of artificial intelligence and decentralized networks — frontier AI infrastructure, crypto-native AI agents, Bittensor subnets, DePIN economies, and tokenized compute.

THE DAILY SIGNAL

The stories that move AI & crypto markets — before the market reacts.

Free. 7am ET. Five stories. 62,400 readers.