Agentic AI – systems that can plan, act, and learn independently – is being adopted faster than traditional and generative AI, yet few organisations have developed the management frameworks to keep pace. In less than two years, 35% of companies are already exploring agentic AI, with another 44% of companies planning to deploy it soon. However, few organisations have developed the management frameworks necessary for redesigning their workflows, governance models, investment planning, and talent strategies to keep up with this unprecedented pace.

These are among the findings of the ninth annual global research study on AI and business strategy released today by MIT Sloan Management Review and Boston Consulting Group (BCG). Titled “The Emerging Agentic Enterprise: How Leaders Must Navigate a New Age of AI,” the report draws from a survey of 2,102 executives across 21 industries and 116 countries, as well as interviews with senior leaders.

Respondents in APAC and Africa are substantially more likely to view AI agents as collaborative colleagues rather than tools—82% of African participants see them as colleagues versus just 18% as tools. This reflects a fundamental shift in how enterprises conceptualise AI, where AI agents’ evolving capabilities continue to blur traditional roles and responsibilities, while organisations struggle to redesign their workflows, governance models, investment planning, and talent strategies.

“Historically, we had a nice, clean separation between technology and people, with management processes designed around that distinction. But agentic AI is neither a tool nor a teammate — it’s both and thrives in that blur. The organisations that will succeed are those that recognise agentic AI’s dual nature as a feature, not a bug,” said report co-author Sam Ransbotham, an analytics professor at Boston College.

Unlike earlier technologies, agentic AI systems are more than just tools to be operated or assistants waiting for instructions. Increasingly, they behave like autonomous teammates, capable of executing multistep processes and adapting as they go. Notably, 76% of the executives surveyed view agentic AI as more like a coworker than a tool.

“Agentic AI has the power to transform entire workflows and challenge existing business processes. The organisations that will succeed are those that put in the effort to reimagine their processes and not just force-fit agentic AI into existing ones,” said Shervin Khodabandeh, BCG managing director and senior partner, a leader of the firm’s AI business and co-author of the report.

Other key findings include:

While 66% of leading agentic AI organisations expect operating model changes compared with 42% of beginners, African respondents demonstrate particularly high expectations: they anticipate AI agents will handle 47% of their current workload today and 60% within three years, among the most optimistic projections globally.
58% of agentic AI leaders expect governance structure changes within three years, with expectations that AI systems will have decision-making authority growing 250%.
43% of agentic AI leaders anticipate being more open or willing to hire generalists in place of specialists; 45% anticipate a reduction in layers of middle management; and 29% expect fewer entry-level roles.
95% of individuals at leading agentic AI organisations report AI positively impacting their job satisfaction.
73% of agentic AI leaders expect that using the technology will change their organisation’s ability to differentiate itself compared with 53% of those with no agentic AI adoption. Additionally, 76% of individuals working at leading agentic AI organisations believe that using agentic AI will affect their ability to differentiate themselves from their co-workers compared with 49% of individuals working in organisations with no agentic AI adoption.

Organisations now face a challenge as a result of agentic AI’s uniqueness as both a tool and a co-worker: managing a single system that demands both human resource approaches and asset management techniques. The report offers evidence-based recommendations on how to proceed.

Download the report here.