At Think 2026, IBM CEO Arvind Krishna said enterprises pulling ahead are redesigning how their business operates, not simply deploying more Artificial Intelligence. Generative Artificial Intelligence has often centered on chatbots, summaries and pilots that have not scaled, while agentic Artificial Intelligence is positioned as a more significant shift because autonomous agents can act, orchestrate workflows, make decisions and hand off work across systems and teams.
The IBM IBV 2025 CEO Study surveyed 2,000 CEOs across 33 countries. The study found that only 25% of Artificial Intelligence initiatives have delivered expected ROI over the last few years. And just 16% have scaled Artificial Intelligence across the enterprise. The data is clear: 85% of CEOs expect Artificial Intelligence efficiency investments to pay off by 2027. IBM argues that the gap is less about technology procurement and more about workforce readiness, including the ability to orchestrate agents, govern Artificial Intelligence systems and redesign work around them.
IBM is tripling its entry-level hires for roles some expected Artificial Intelligence to eliminate, while rewriting job descriptions rather than removing early-career positions. A junior software developer now spends less time on routine coding and more time with clients, while an entry-level HR employee intervenes when chatbots fall short, corrects Artificial Intelligence outputs and escalates issues to managers. Krishna’s view is that when people become more productive, organizations do not shrink, they grow.
IT organizations are being asked to build new capabilities around agent orchestration, multi-agent LLM systems, vector databases, edge Artificial Intelligence deployment, prompt engineering, context design and governance as code. At level 4 of the agentic SDLC framework, governance must be embedded as policy-as-code, not bolted on after the fact. Several roles now exist that did not exist two years ago: the agent governance lead, the head of Artificial Intelligence platform and the agent product owner.
According to the IBM IBV study, 54% of CEOs are now hiring for Artificial Intelligence-related roles that did not exist a year ago. Nearly one-third of the workforce will need to reskill in the next three years to remain competitive. IBM’s position is that enterprises need a skills development machine, not just a hiring machine, to move agentic Artificial Intelligence from experimentation into measurable business outcomes.
