The state of AI in 2025: Agents, innovation, and transformation

This document presents the results of a global survey of nearly 2,000 professionals on the use of artificial intelligence in their organizations during 2025. The report offers a complete overview of the current state of AI in the business world.

Main findings: Although AI use has become widespread (88% of organizations use it in at least one function), two-thirds have not yet begun to scale it at enterprise level. Most remain in experimentation or pilot project phases, without yet achieving tangible benefits across the entire organization.

AI agents: The document explores the emerging field of AI agents (systems capable of planning and executing multiple steps autonomously). 62% of organizations are experimenting with these systems, although their widespread implementation remains limited, with greater adoption in areas such as IT and knowledge management.

Economic impact: Only 39% report impact on EBIT (operating profit), but clear benefits are observed in innovation (64%), customer satisfaction, and cost reduction in specific functions. The most successful companies, called "high performers," are distinguished by a transformative vision: they seek not only efficiency, but also growth and innovation.

Success factors: The report identifies key practices that differentiate leading organizations: fundamental redesign of work processes (not just automation), active commitment from senior leaders, greater investment in AI (more than 20% of digital budget), and establishment of robust processes for human validation of results.

Labor impact: Expectations about employment are diverse: 32% anticipate workforce reduction, 43% no change, and 13% increase. At the same time, demand for specialized AI roles (software engineers, data scientists, machine learning engineers) continues to grow, especially in large companies.

Industry trends: AI and agent use is more pronounced in technology, media/telecommunications, and healthcare, while sectors like construction and pharmaceutical products show lower penetration. The most common use cases include contact center automation, deep research in knowledge management, support in creating marketing content, and information processing through conversational interfaces.

Risk management: The document shows that organizations are taking AI risks more seriously. Result inaccuracy is the most commonly experienced risk, with 30% reporting negative consequences, followed by explainability, privacy, and regulatory compliance issues.

Conclusions and recommendations: The report emphasizes that overcoming the pilot phase requires more than implementing tools: organizations must invest in employee training (upskilling), establish solid governance structures, redesign complete processes rather than just automating tasks, and maintain consistent leadership commitment. The difference between those who obtain real value and those who don't lies in treating AI as a transformation catalyst, not as a simple efficiency tool.

This report is especially valuable for executives, digital transformation managers, and professionals seeking to understand how organizations globally are adopting AI, what separates leaders from the rest, and what challenges they face on the path to AI maturity.

Key points

• Widespread adoption but not scaling: 88% of organizations use AI, but two-thirds remain in experimental or pilot phase.
• Limited impact on results: Only 39% report impact on enterprise-wide EBIT, although there are benefits in specific functions.
• Leaders think about transformation: Successful companies seek innovation and growth, not just operational efficiency.
• Redesigning processes is critical: 55% of high performers redesign complete workflows, not just automate tasks.
• AI agents growing: 62% experiment with autonomous agents, mainly in IT and knowledge management.
• Differential investment: Leaders allocate more than 20% of their digital budget to AI, versus 10% for others.
• Committed leadership makes the difference: High performers are 3 times more likely to have active support from executives.
• Diverse labor expectations: 32% anticipate workforce reduction, 43% no change, 13% increase due to AI.
• Risks materializing: 30% have experienced negative consequences from AI result inaccuracy.
• Leading sectors: Technology, telecommunications, and healthcare show greater adoption of AI and agents.
• Key recommendations: Invest in employee training, establish robust governance, and redesign complete processes.