Companies are committing substantially more budget to employee AI training than to agentic AI deployment itself, with 39% planning training investments over the next 12 months compared to 21% for AI agents. This inversion reveals a critical gap between technology readiness and organisational capability. Nearly 30% of enterprises already run AI agents in production whilst a further 26% pilot them, yet the majority recognise that implementation velocity has outpaced workforce readiness. The Liferay survey identifies the core issue: employees lack not just familiarity with AI tools, but the conceptual framework to deploy them effectively. Data privacy and employee trust remain the stated concerns, but the emphasis on training suggests companies understand the real bottleneck is behavioural and cognitive rather than technical.
The shift from copilots to agentic systems fundamentally changes what employees must learn. Where copilot adoption required prompt engineering and question-asking, agentic AI demands a different competency entirely: the ability to determine when autonomous systems should act versus when human oversight is necessary. This represents a move from tool usage to decision governance, and it explains why critical thinking course enrollments have surged 185% year-on-year. For CX teams specifically, this creates both opportunity and risk. Support leaders deploying Zendesk or Freshdesk with agentic capabilities must now invest in training that teaches agents to review and validate AI decisions rather than simply prompt them—a fundamentally different skill set. The question becomes whether your current training infrastructure can support this shift, or whether you're deploying agents into teams still operating under copilot-era mental models.
The data also exposes a vendor challenge. If 39% of organisations prioritise training investment, then platforms that bundle robust training, decision-review workflows, and human approval mechanisms into their core offering will capture disproportionate value. Zendesk's outcome-based pricing model and emphasis on balancing AI with human touch align directly with this market signal. Conversely, vendors selling pure agentic capability without training scaffolding may find adoption stalls at the pilot stage—which is precisely where 26% of companies currently sit. For support leaders, this means evaluating not just agent capability but the training ecosystem your platform provides.
With agentic AI rollouts and trials more common, companies plan to invest in employee training, Liferay survey finds.