Data privacy, cost & legacy barriers slow enterprise AI

Research from Cloudera has identified data privacy, legacy system integration, and high implementation costs as the main barriers to deploying agentic artificial intelligence in enterprises worldwide.

The findings are drawn from a survey of nearly 1,500 enterprise IT leaders across 14 countries, commissioned by Cloudera for its latest report, "The Future of Enterprise AI Agents." The survey shows that 96% of respondents plan to increase the use of AI agents in the next year, and half intend for this expansion to be significant and organisation-wide.

Despite the enthusiasm, data privacy concerns were cited as the leading challenge, mentioned by 53% of respondents. Integration with legacy systems followed at 40%, and high implementation costs were noted by 39%. The report suggests that these barriers are linked by a shared requirement for robust, unified data management and governance.

In terms of adoption, a majority (57%) of enterprise IT leaders reported having implemented AI agents in the past two years, with 21% doing so in just the last year. Cloudera notes this as evidence of rapid momentum in the field, which is expected to continue growing.

The research also reveals the approaches organisations are taking to deployment. Two-thirds (66%) of respondents indicated they are building agents on enterprise AI infrastructure platforms. Meanwhile, 60% are making use of agentic capabilities embedded in existing core applications. Cloudera interprets this as a preference for scalable, secure, and "close-to-data" deployments, reflecting a hybrid approach currently favoured by many enterprises.

When asked about the strategic importance of AI agents, 83% of the organisations surveyed stated that investing in such technologies is essential to maintaining competitiveness in the market. The report lists performance optimisation bots (66%), security monitoring agents (63%), and development assistants (62%) as the leading applications for these deployments.

Some of the identified benefits of effective agentic AI implementation include operational agility, cost savings, and improved customer engagement. However, realising these outcomes remains a challenge for many organisations as they seek to overcome the aforementioned barriers.

Cloudera's research outlines suggested paths forward for companies seeking to begin or expand their use of AI agents. Among the recommendations is starting with contained, high-impact projects, such as internal IT support agents. These projects can provide rapid returns on investment and help build support for broader deployments across a business.

Abhas Ricky, Chief Strategy Officer at Cloudera, said: "AI agents have moved beyond experimentation—they're now delivering real automation, efficiency, and business results. We're seeing enterprises run hundreds of models in production, all demanding high-fidelity, well-managed data to drive better outcomes."

Ricky continued: "In 2025, agentic AI is taking center stage, building on the momentum of generative AI but with even greater operational impact. Cloudera is enabling this transformation through the world's first Enterprise AI Ecosystem, helping global organizations design secure, scalable, and integrated AI workflows that turn data into action."

The report details sector-specific uses of AI agents. In finance and insurance, the main applications are fraud detection (56%), risk assessment (44%), and investment advisory (38%), with agents being used to flag suspicious transactions and simulate market scenarios. In manufacturing, organisations are integrating agents for process automation (49%), supply chain optimisation (48%), and quality control (47%). Healthcare uses include appointment scheduling (51%), diagnostic assistance (50%), and medical records processing (47%). In telecommunications, top use cases are customer support bots (49%), customer experience agents (44%), and security monitoring agents (49%).

The survey data indicates a fast-expanding role for agentic AI, though persistent challenges around data privacy, integration, and cost are prompting organisations to focus on strong data management and targeted initial deployments as they advance their use of AI agents.