An overwhelming majority of IT leaders, at 87%, consider investment in AI agents essential for maintaining competitiveness, according to a survey conducted by Cloudera. The poll, which included responses from 1,484 IT leaders spread over 14 countries, underscores the increasing popularity of agentic AI.
Findings from the survey reveal that 96% of respondents plan to expand their use of AI agents within the next year. Half of these enterprises aim for widespread, enterprise-level implementation, indicating a significant acceleration in the adoption of agentic AI. Notably, 57% reported initiating implementation within the past two years, with 21% starting only in the last year.
Enterprises are prioritising investment in specific areas such as performance optimisation bots (66%), security monitoring agents (63%), and development assistants (62%). These tools are intended to enhance productivity and resilience across various operations. To support this adoption, organisations are utilising enterprise AI infrastructure platforms (66%) and integrating agent capabilities within existing core applications (60%). However, this rapid uptake necessitates a robust and scalable data infrastructure to avoid falling behind, said Cloudera.
Once implemented, AI agents appear to deliver substantial benefits to host businesses. The survey identified improvements in existing generative AI models (81%), customer support applications (78%), process automation (71%), and predictive analytics (57%) as key advantages. Currently, IT operations see the most deployment of AI agents (61%), with additional applications in customer support (18%) and marketing (6%).
Challenges in adopting agentic AI
Despite these benefits, challenges remain in adopting agentic AI. Data privacy is a major concern for 53% of IT leaders. Other significant challenges include integration with legacy systems (40%) and implementation costs (39%). Enterprises face the complex task of protecting sensitive data while ensuring its effective use across the AI lifecycle.
Concerns about trust and bias are also prevalent as AI agents assume more responsibility. The survey found that 51% of enterprise leaders express worries about bias in AI systems. In response, organisations are implementing measures such as human reviews, diversified training data, and formal fairness audits to address these issues. However, 14% of respondents acknowledge taking only minimal steps to combat bias.
A recent survey by PagerDuty shows that 51% of companies have integrated AI agents into their workflows, with another 35% planning to follow suit in the next two years. The survey, which involved 1,000 IT and business executives from the US, UK, Australia, and Japan, underscores a trend towards quicker integration of agentic AI, with 55% of businesses keen to fast-track deployment.
Meanwhile, a Gartner study forecasts that by 2029, agentic AI will manage 80% of standard customer service inquiries independently, potentially reducing operational costs by 30%. Unlike earlier AI models focused on text generation, agentic AI is designed for task execution without human aid. Initial applications are expected in customer service, a sector characterised by routine tasks challenging for traditional machine learning.
Read more: Agentic AI to automate 80% of customer service queries by 2029, predicts Gartner
