Corporate investment in AI-driven customer support continues to accelerate, yet organisations face two critical legal exposures that demand immediate attention from CX leadership. The first centres on intellectual property and data ownership—when AI systems are trained on proprietary customer data, support interactions, or internal knowledge bases, the resulting models and outputs exist in a legal grey zone. Vendors like Zendesk and Salesforce now embed generative AI directly into their platforms, but the question of who owns the trained models, whether customer data becomes training material for broader systems, and how to protect competitive advantage through support interactions remains unresolved in most jurisdictions. The second risk involves liability and accountability: when an AI agent makes a decision, provides incorrect information, or escalates a case inappropriately, determining fault becomes complex. Is responsibility with the vendor, the organisation deploying the system, or the support team member who failed to catch the error? This ambiguity intensifies when AI recommendations influence high-stakes decisions or when systems operate with minimal human oversight.
For CX teams already operating within these platforms, the implications are substantial. Teams implementing Agentforce, Freshdesk's AI features, or similar solutions must establish clear governance frameworks around what data feeds these systems and how outputs are monitored before reaching customers. The trend of companies bringing humans back after betting big on AI reflects not just performance concerns but also risk mitigation—human-in-the-loop approaches provide an accountability layer that pure automation cannot. Support leaders should audit their current deployments against vendor terms of service, particularly around data retention and model training, and establish escalation protocols that create an auditable trail when AI recommendations are overridden or questioned. Without this groundwork, organisations risk both regulatory exposure and the erosion of customer trust if AI-generated errors surface publicly.
AI adoption is a corporate imperative but it comes with two significant legal risks Financial Post