The rise of AI agents in 2025 marks a strategic milestone in the digital transformation of companies. Previously limited to small-scale experiments, agentic AI is now set to establish itself as a driver of efficiency, automation, and operational innovation. According to a survey conducted by Cloudera with 1,484 IT leaders across 14 countries, 96% of organizations plan to increase their use of AI agents over the next 12 months.
Massive, yet recent adoption
Although recent, the adoption of AI agents is progressing rapidly: 57% of companies have introduced them in the past two years. The investment is significant, with 83% of executives considering it essential to invest in agentic AI to remain competitive.
Their investment priorities focus on:
- Performance optimization (66% of companies);
- Cybersecurity and threat monitoring (63%);
- Software development (62%).
Companies favor two complementary strategies:
- 66% develop their agents on dedicated AI infrastructure platforms, ensuring security and scalability.
- 60% prefer integration into their existing business applications, facilitating implementation without deep restructuring.
This hybrid model reflects a desire to adopt AI progressively while ensuring that agents evolve close to the company's strategic data.
Opportunities and challenges
Perceived benefits are numerous: improvement of existing GenAI model performances (81%), better customer support (78%), enhanced automation (71%), and more refined predictive analysis (57%). This potential is not limited to IT — although it is the main area of deployment (61%) — but is gradually extending to marketing (6%) and customer service (18%).
However, several barriers slow this adoption. Three concerns consistently emerge:
- Data privacy (53% of respondents);
- Integration with existing systems (40%);
- Implementation costs (39%).
Use cases by sector
Agentic AI adapts to the specifics of each industry sector. The main use cases identified by the study are as follows:
- Finance & Insurance: Fraud detection (56%), risk assessment (44%), and investment advice (38%);
- Manufacturing: Process automation (49%), supply chain optimization (48%), and quality control (47%);
- Healthcare: Appointment scheduling (51%), diagnostic assistance (50%), and medical record processing (47%);
- Telecommunications: Customer support bots (49%), customer experience agents (44%), and security monitoring agents (49%).
Equity and responsibility, key challenges for the future
As AI agents are entrusted with increasingly strategic tasks, issues related to responsibility, equity, and transparency become major priorities. More than half (51%) of the executives surveyed express significant concerns about potential biases in these systems. In response, 38% of companies are implementing equity audits and diversifying their training data sets, while 36% introduce regular verification mechanisms.
However, nearly 14% of organizations have undertaken only minimal or occasional actions to address these issues, illustrating a heterogeneity in the ethical maturity of practices.
According to Cloudera, successful adoption relies on optimal data management and a progressive approach, starting with high-impact projects such as IT support agents. Their return on investment will help establish a solid climate of trust internally and lay the groundwork for large-scale deployments.
To better understand
What is agentic AI and how does it differ from other types of artificial intelligence?
Agentic AI refers to autonomous systems that can perceive their environment, reason, and act independently to achieve defined goals without constant human intervention, differing from more passive AI requiring direct instructions.
What are the data protection regulations that could influence the deployment of AI agents in companies?
Regulations such as GDPR in Europe impose strict data management constraints, affecting the use of AI agents that must ensure privacy, limit unauthorized access, and ensure compliance with user consent.