AI agents are enabling a new realm of productivity and efficiency. These assistants not only generate content but also take action, which is why enterprises are looking to agentic AI to reimagine workflows, productivity, and innovation. But what does that look like in practice, and where should enterprises focus their agentic aspirations? Cloudera’s recent global survey of enterprise IT leaders highlights how adoption is accelerating and shaping enterprise priorities.
Survey respondents emphasised that adoption of agentic AI is urgent, that the ROI can justify the effort and spending, that infrastructure is key to enabling AI agents, and that model bias is still a problem.
The future of enterprise AI is agentic
Interactive agentic systems are already reasoning, planning, and collaborating with users in new ways. Many enterprise leaders see adoption as a strategic priority. According to Cloudera’s survey, 89% of IT leaders plan to increase their use of agents in the next 12 months. More than half of them (56%) are aiming for widespread, enterprise-level implementation.
For many, this uptake is relatively recent, with 25% of respondents saying their organisations began implementing agents within the last two years.
When it comes to business functions, organisations are primarily embedding AI agents in IT operations (62%), followed by finance (15%) and customer support (14%). The vast majority (87%) reported that their prior investments in generative AI prepared them well to implement AI agents, which is encouraging for those just beginning to experiment.
As adoption expands, enterprises must assess whether their infrastructure is ready to fully support AI agents.
AI agents and ROI
As enterprises explore where agentic AI can deliver the most impact, early use cases are already demonstrating value. According to the survey, 81% of respondents said they have seen tangible benefits.
AI agents are being deployed for customer support (77%), process automation (56%), and predictive analytics (44%), showing that many organisations begin with well-defined, ROI-driven domains. These areas provide opportunities to deliver results through automation, whether in IT helpdesks or in leveraging predictive analytics to stay ahead of cyberattacks.
Agentic AI adoption has been catalysed by advancements in generative AI. While generative AI assistants can increase efficiency and take on tasks individuals may not have time for, the benefits often accrue at a limited scale. Agentic AI extends these advantages across the enterprise. For example, vendor contract summarisation and clause management can be expanded to partner and customer contracts, and models trained on core marketing messaging can be made available to all market-facing contributors.
Additionally, agents can leverage specialised models to plan and orchestrate tasks and reason through complex challenges before selecting the most appropriate actions. As APIs are integrated into the agentic framework, agents will be able to take on an increasing number of day-to-day tasks across workflows within the organisation.
Addressing bias concerns
With greater autonomy comes an increased need for accountability, which remains top of mind for IT leaders considering agentic AI. In Cloudera’s survey, 71% of enterprise leaders reported significant concerns about bias in AI systems.
As AI agents take on mission-critical tasks, enterprises are working to establish accountability and proper governance. A sizeable number of respondents (45%) said they are implementing processes that include human reviews, diversified training data, and formal fairness audits, with another 35% having introduced some bias-check measures.
Data quality and availability also remain significant challenges in AI implementations of any kind. The strength and flexibility of enterprise data infrastructure will be essential in addressing these issues.
Deploying AI agents securely and at scale
Enterprises are moving quickly to turn their agentic AI ambitions into enterprise-grade applications with measurable outcomes. To achieve this, they must focus on building secure, scalable environments that support the deployment of AI agents.
This means ensuring that data governance is robust, that infrastructure can handle increased workloads, and that models are trained and monitored with fairness and accountability in mind. As adoption expands across functions, enterprises that align their AI agent strategies with existing security frameworks and compliance requirements will be best positioned to deploy these systems responsibly.
By grounding adoption in strong data practices, infrastructure readiness, and clear governance, enterprises can scale AI agents with confidence and ensure they deliver value in a secure and sustainable way.
