The Breakout Year for Enterprise AI Agents

2025 is shaping up to be a landmark year for enterprise AI. Advancements in generative AI (GenAI) and large language models (LLMs) have brought the transformative power of agentic AI to the forefront of every IT leader's mind. While often mistaken for simple chatbots, AI agents are far more advanced - autonomous tools capable of executing complex, goal-oriented tasks. Their impact is already felt across sectors - from real-time fraud detection in finance to workflow optimization in manufacturing and precision diagnostics in healthcare.

With investment and adoption of AI agents soaring, how are enterprise leaders prioritizing the technology in their own organizations? To explore how organizations are embracing this new wave of AI, Cloudera surveyed 1,484 enterprise IT leaders across 14 countries. The findings reveal not just a strong commitment to AI agents, but a transformative shift in how businesses are planning, deploying, and evaluating them.

Enterprises Are Doubling Down on Agentic AI Investment
Adoption of AI agents is no longer an experimental endeavor—it’s a strategic imperative. An overwhelming 87% of respondents said investing in AI agents is essential to stay competitive. Even more telling, 96% plan to increase their use of agents in the next 12 months, with half of them aiming for widespread, enterprise-level implementation .

Considering the uptick in investment, agentic AI adoption is actually a relatively new development for many enterprises. In fact, a majority (57%) of surveyed respondents said their organizations only began implementing them in the last two years, with 21% starting within the past year . This rapid embrace of agentic AI is reflected in how organizations are prioritizing their investments.

Top areas of investment reflect this mindset. Organizations are prioritizing performance optimization bots (66%), security monitoring agents (63%), and development assistants (62%)—tools that promise to enhance both productivity and resiliency . So, how are organizations enabling those AI agents to take hold? According to survey respondents, 66% said they are using enterprise AI infrastructure platforms to develop and deploy AI agents. And 60% are taking advantage of agent capabilities embedded within their existing core applications.

As adoption accelerates, these trends signal a critical need for enterprises to have a reliable, scalable, data infrastructure in place. Given the universal commitment to investing in that adoption, organizations need to ensure that infrastructure is in place quickly, or risk getting left behind on the road to agentic AI.

Where AI Agents Deliver Value—and What’s Holding Them Back
Once successfully implemented, AI agents can deliver tremendous value to organizations. Some of the tangible benefits that enterprises are seeing include improving existing GenAI models (81%), and applications that include customer support (78%), process automation (71%), and predictive analytics (57%) .

For most enterprises, those AI agents are most deeply embedded in their IT operations (61%). This was followed by customer support (18%) and marketing (6%), as areas of implementation. Companies that have incorporated agents into IT functions are more likely to branch out into customer and marketing use cases, suggesting that IT is the natural launching pad for wider agentic AI integration .

The benefits are clear, and the use cases associated with AI agents have serious potential to transform the way enterprises function. But the path forward still comes with some apprehension. Specifically, over half of IT leaders identified data privacy (53%) as a concern with adopting agentic AI, while integration with legacy systems (40%) and implementation costs (39%) followed close behind .

Organizations have to manage a delicate balance of protecting sensitive data but also ensuring it is leveraged throughout the AI lifecycle. Any inadvertent exposure of that sensitive data could hamper the quality of AI outputs and leave organizations out of compliance with key regulations, like DORA.

AI Built with Accountability
The concerns many organizations have often boil down to a few major considerations—trust and bias. As AI agents gain more responsibility and take control over mission-critical tasks, questions of accountability, fairness, and transparency are becoming top of mind. Over half (51%) of enterprise leaders reported significant concerns about bias in AI systems .

Understanding of bias is growing, and enterprises are taking added steps to build in accountability and govern AI properly. A sizable number of respondents (38%) are implementing a number of processes that include human reviews, diversified training data, and formal fairness audits. Beyond those steps, another 36% said they have introduced some bias-check measures, like periodic human reviews or bias-detection tools.

But for all the mitigation efforts that are under way, a gap still exists with some enterprises. Cloudera found that 14% of respondents said they have only taken minimal or ad hoc steps so far to combat bias. AI agents cannot function without accountability and fairness. Enterprises adopting this technology need to ensure that they are taking the necessary steps to reduce the impact bias can have on real-world outputs.

Accelerating AI Agent Adoption
2025 is poised to be a massive year, with accelerating adoption of AI agents unlocking new use cases for businesses everywhere. And as investment in the technology accelerates, prioritizing adoption of AI agents is quickly becoming nonnegotiable for long-term success.

Find out what else Cloudera’s survey uncovered and take a deeper dive into the state of AI agents in 2025.