AI adoption in customer service has moved decisively beyond experimentation into operational infrastructure. Salesforce's report documents a sharp acceleration: AI usage across service centres reached 85%, whilst reliance on agentic systems jumped from 39% to 66% in a single year. The shift reflects a fundamental change in capability—these systems now execute transactions autonomously rather than merely suggest answers. AI agents modify invoices, process requests, and escalate cases with human-ready summaries without human intervention in 40% of interactions. Most organisations deploying these systems have rolled them across five or more channels simultaneously, embedding AI as a core operational layer rather than a siloed experiment. Within two months of deployment, 70% of organisations reported measurable gains in satisfaction, response speed, and agent productivity, with staff particularly noting relief from application-switching overhead.
Yet this expansion masks a critical infrastructure problem. Approximately 60% of service leaders report dissatisfaction with their data readiness, and the challenge compounds as customer records fragment across unintegrated platforms—precisely the scenario many teams face when layering agentic AI onto legacy Zendesk, Freshdesk, or Salesforce implementations without prior data consolidation. The trust gap presents a second friction point: whilst two-thirds of executives believe customers trust AI systems, fewer than half of actual customers express confidence in AI handling their issues, particularly in sensitive or complex scenarios. This divergence suggests that organisations deploying agentic AI without clear escalation pathways or transparency mechanisms risk customer satisfaction gains eroding once systems encounter edge cases.
The role of human agents is recalibrating rather than disappearing. Workers are shifting from routine query resolution to complex case management, system decision review, and genuine exception handling—a transition that demands different hiring, training, and retention strategies than traditional support models. What remains unclear is whether organisations currently understaffed for complex work, or those with weak data governance, can realistically manage this transition without either degrading service quality or incurring significant restructuring costs. The vendors consolidating agentic AI capabilities—as evidenced by Digital Workforce's acquisition of Front AI's customer service business—are betting that integrated platforms will become table stakes, which may pressure mid-market teams to choose between expensive platform migrations or accepting fragmented AI deployments.
Report: AI adoption expands in customer service centers جولة