It’s clear that AI agents are enabling a new realm of productivity and efficiency gains – AI assistants that not only generate content but take action – which is why enterprises are looking to agentic AI to reimagine workflows, productivity, and innovation. But what will that look like in practice, and where should enterprises focus their agentic aspirations? Cloudera’s recent global survey of enterprise IT leaders suggests that 2025 is the year we find out. 

Survey respondents believe that the adoption of agentic AI is urgent, that the ROI will justify the effort and spend, that infrastructure is key to enabling AI agents, and that model bias is still a problem. Let’s dive into the details.

The future of enterprise AI is agentic, and the clock is ticking

Interactive agentic systems are already reasoning, planning, and collaborating with users in new ways. The adoption of AI agents is no longer optional, it’s a strategic necessity. According to Cloudera’s “The Future of Enterprise AI Agents” survey, 96% of IT leaders plan to increase their use of agents in the next 12 months. Even more telling: half of them are aiming for widespread, enterprise-level implementation​.

This uptick in AI adoption is relatively new for many enterprises. A majority (57%) of surveyed respondents said their organizations began implementing agents within the last two years. 

Primarily, organizations are deploying performance optimization bots (66%), security monitoring agents (63%), and coding assistants (62%) – tools focused on enhancing both productivity and resiliency in IT and customer service. The vast majority (85%) report that their prior investments in GenAI prepared them well to implement AI agents, which is encouraging for those who are just beginning to experiment.

As adoption accelerates, enterprises must assess whether their infrastructure is ready to fully enable AI agents. Sixty-six percent of survey respondents are using enterprise AI infrastructure platforms to develop and deploy AI agents, with 60% taking advantage of agentic capabilities embedded within their existing core applications.

AI agents will significantly increase ROI

As enterprises explore where agentic AI can deliver the most impact, early use cases are already demonstrating value. Eighty-one percent of survey respondents have already seen tangible benefits.

AI agents are being used for customer support (78%), process automation (71%), and predictive analytics (57%)​ – showing that many companies start adoption in well-defined, ROI-driven domains. These areas provide ample opportunity to drive tangible results via automation, whether deploying IT helpdesk agents or leveraging predictive analytics to stay ahead of cyber attacks. As teams deploy and scale these AI agents, having low-code and no-code tools will be critical to ensure their success along the way.

Ninety-eight percent of surveyed organizations are either already using agentic AI to orchestrate GenAI use cases or plan to do so in the near future. It is the tight coupling between GenAI capabilities and agentic AI applications that holds the key to unleashing substantial ROI. 

GenAI assistants improve human resource efficiency and increase departmental capability by taking on tasks humans couldn’t make the time for. But those advantages often benefit only a few individuals. Agentic AI is a force multiplier. It extends the capabilities and benefits across the enterprise. For example, vendor contract summarization and clause management can be extended to partner and customer contracts; models trained on core marketing messaging can be made available to all market-facing contributors.

Additionally, agents leverage specialized LLMs to plan and orchestrate tasks and reason through complex challenges before selecting the most appropriate actions. As APIs are integrated into the agentic framework using low-code/no-code tools, agents will be able to perform an increasing number of day-to-day tasks in workflows throughout the organization.

That’s when ROI will skyrocket.

Bias concerns must be addressed

With greater autonomy comes an increased need for accountability, which is top of mind for IT leaders considering agentic AI. In Cloudera’s survey, over half (51%) of enterprise leaders reported significant concerns about bias in AI systems​.

As AI agents take control over mission-critical tasks, enterprises are working to establish accountability and proper governance. A sizable number of respondents (38%) are implementing processes that include human reviews, diversified training data, and formal fairness audits, with another 36% having introduced some bias-check measures.

Data quality and availability issues remain significant technical challenges in AI implementations of any kind. The solution lies in the strength and flexibility of enterprise data infrastructure. 

Enlisting a trusted partner

Cloudera is helping organizations turn their agentic AI ambitions into enterprise-grade applications with substantial business outcomes. Our Enterprise AI platform combines trusted data infrastructure with scalable AI development tools, including low-code/no-code capabilities with AI Inference services and AI studios, which facilitate the safe deployment of AI agents at scale.

Cloudera accelerates the pathway from experimentation to production and facilitates the embedding of cloud-native models inside private, highly secure data estates, effectively reducing the risk of AI to the current risk exposure of data environments.

AI agents will be implemented at an even faster pace than generative AI assistants, so 2025 is the year to act. Those with the right tools and partners will become outcome-generators, and distance themselves from the out of touch.

 About the Author:

Shayde Christian is Chief Data and Analytics Officer at Cloudera. He guides data-driven cultural change for Cloudera to generate maximum value from data. Christian enables customers to get the absolute best from their Cloudera products such that they can generate high-value use cases for competitive advantage.

Previously a principal consultant, Christian formulated data strategy for Fortune 500 clients and designed, constructed, or turned around failing enterprise information management organizations. He enjoys laughter and is often the cause of it.