Agentic AI: Expectations, Readiness, Results

Agentic AI —capable of executing complex tasks autonomously, making decisions, and coordinating actions without constant human intervention— is redefining how organizations operate and compete. This Harvard Business Review Analytic Services report, sponsored by AWS and based on a survey of 623 executives and technology decision-makers conducted in July 2025, offers a detailed picture of how companies are approaching this transformation: what they expect from the technology, how prepared they are to adopt it, and what results early adopters are already achieving.

The data reveals widespread enthusiasm: 90% of respondents expect most organizations in their industry to use agentic AI in the future, and 84% agree it will transform their business. Yet this optimism contrasts with a more complex reality. Only 26% of organizations consider themselves truly effective at using any type of AI to generate positive business outcomes. And just 5% have well-defined metrics to measure the success of their initiatives —one of the most critical blind spots identified in the report.

The document distinguishes three adoption profiles: leaders (34%), who are already using agentic AI widely or in specific use cases; followers (40%), in pilot or planning stages; and laggards (26%), who have yet to take formal steps. From this classification, the report analyzes what sets the most advanced organizations apart and what practices can guide others.

One of the report's central messages is that agentic AI should not be seen as a tool to do the same things faster, but as an opportunity to rethink processes and business models from the ground up. This distinction between automation and real transformation recurs throughout the executives' testimonies, and marks the difference between projects that generate value and those that simply make existing processes more expensive.

The report examines four key areas of organizational readiness: data architecture, governance, human capital, and strategy. In all of them, most organizations remain only partially prepared. Governance takes on particular importance given the autonomous nature of agentic AI, which raises new ethical, legal, and operational challenges. The human factor emerges as the hardest barrier to overcome: the lack of specialized talent (48%) and resistance to change are obstacles as significant as technological ones, and change management proves to be a decisive element for successful implementation.

Despite these challenges, organizations that have moved ahead report tangible benefits: greater organizational productivity (36%), better data-driven decision making (35%), and cost savings (33%). The report also gathers best practices from experts and executives at companies such as Syngenta, Vanguard, Experity, and McAfee, with concrete guidance on data quality, governance structures, AI literacy, and alignment between technology and business strategy.

Aimed at executives, technology leaders, and business decision-makers, this report is an essential reference for those seeking to make informed decisions about how to incorporate agentic AI responsibly, strategically, and with real business impact.

Key points

• 90% expect widespread agentic AI adoption; 84% believe it will transform their business
• Only 26% of organizations are effective at using AI to generate real business outcomes
• The key is not automating existing processes, but transforming them from the ground up
• Three adoption profiles: leaders (34%), followers (40%), and laggards (26%)
• Main barriers: lack of talent (48%) and absence of a clear roadmap (46%)
• Readiness in data, governance, and human capital remains partial in most organizations
• Change management and the human factor are as critical as technology
• Only 5% have well-defined metrics to measure the success of their initiatives
• Early results: greater productivity (36%), better decisions (35%), and cost savings (33%)
• Aligning AI strategy with business strategy is the key differentiator for leading organizations