Cresta's launch of Training Simulator addresses a critical infrastructure failure in contact center AI deployment: the systematic underpreparing of agents for the conversations that matter most. Whilst 79% of service professionals are investing in agentic AI and contact center automation is accelerating, a stark gap persists between organizational AI capability and agent readiness. Zendesk's own data exposes this acutely—72% of CX leaders claim adequate generative AI training exists, yet 55% of agents report receiving none. The problem is not technological but structural. Traditional training infrastructure—scripted role-plays, static eLearning, supervisor-led sessions—was designed for onboarding, not continuous upskilling at scale. As routine interactions shift to automation, the calls that reach human agents become exponentially more complex: high-stakes retention saves, commercial upsells, escalations laden with customer frustration. Agents face these conversations with minimal practice, producing avoidable failures on the interactions that directly impact revenue and retention.
Cresta's simulator approach is methodologically sound: it generates AI-powered simulated customers from real conversation data, responds dynamically to agent inputs, and grades performance against live quality criteria. This grounds training in organizational reality rather than generic scenarios. The implications for CX teams are immediate. For Zendesk administrators and support leads already managing agent assist deployments, this raises a pointed question: are your quality frameworks and coaching systems equipped to identify and close readiness gaps before agents encounter live customers? The data suggests most are not. Gartner identifies people and change management as the leading cause of enterprise AI project failures, yet contact center teams routinely treat AI upskilling as a one-time event rather than a continuous performance system. Organizations that treat agent readiness as infrastructure—not afterthought—will see faster ramp times, lower attrition, and measurably stronger outcomes on revenue-critical calls.
The broader implication is that contact center AI underperformance is not a capability problem; it is a will problem. The technology exists. The business case is clear. What remains is whether CX leaders will invest in the training infrastructure required to make AI deployment actually land, or continue accepting preventable failures on the conversations that move the business.
Cresta Targets Contact Center AI Deployment Gap CX Today