Poor AI Customer Bots Pose Growing Brand Risk

Research from customer communications platform Sinch reveals a troubling paradox for brands racing to deploy AI customer service bots. A full 74% of enterprises have already been forced to roll back an AI agent due to governance failures. Even more surprising, companies with the most mature safety protocols and compliance guardrails experienced an even higher rollback rate of 81%. This suggests that heavy investment in oversight isn’t preventing failure; it may actually be a symptom of deeper systemic issues.

The cautionary tale of Air Canada’s chatbot, which invented a nonexistent bereavement refund policy, is now a landmark case. The airline’s attempt to blame the bot as a “separate legal entity” failed in court, and it was ordered to pay damages. But this incident is far from isolated. A car dealership’s bot agreed to sell a Chevy Tahoe for $1 after a prank. A coding startup’s AI support agent fabricated a login policy, sparking mass cancellations. A delivery company’s bot swore at a customer and wrote a poem insulting its own employer. Each viral failure erodes brand trust and customer loyalty.

The guardrail tax is a major culprit. Engineering teams are spending most of their time building and maintaining safety systems instead of improving customer experience. “If governance was the fix, the most mature teams would roll back less, not more,” said Daniel Morris, chief product officer at Sinch. “That’s the guardrail tax that slows organizations down.”

For marketing teams, this tax has a direct cost. Every hour spent rebuilding safety infrastructure is an hour not spent on personalization, channel expansion, or campaign optimization. The Sinch study, which surveyed 2,527 enterprise decision-makers across 10 countries and six industries, found that 84% of teams spend at least half their engineering time rebuilding safety infrastructure from scratch.

The impact of a bot failure lands hardest in two areas: 35% on the support queue and 34% on brand perception. The latter is far harder to repair. Yet the pressure to deploy remains intense. 62% of enterprises already have AI communications agents in production, and 88% expect to deploy one within 12 months.

Infrastructure quality emerged as the single strongest predictor of deployment success, outweighing model choice, team size, and budget. Yet most organizations say their current provider falls short in at least one critical area. Jayashree Iyangar, global lead of CX data and AI at HGS, said the findings match her field observations. “The key question is how AI can be orchestrated seamlessly across multiple channels, not whether it can be deployed in one,” she said. She emphasized that human-in-the-loop oversight remains essential in high-risk service environments, where a bot mishandling a billing issue can cause far more damage than a marketing bot fumbling a promotion.

For marketing teams, the study points to three practical moves. First, prioritize infrastructure quality over model choice or team size when selecting an AI provider. Second, centralize AI governance to handle trust, compliance, and security separately from use-case development. Third, invest in seamless channel orchestration to ensure consistent, safe customer interactions across websites, voice agents, SMS, and email. The brands that get this right will build trust and revenue. Those that don’t risk becoming the next cautionary tale.