**London**: A recent ZDNet survey reveals that 93% of CIOs plan to implement AI agents within the next two years, addressing the challenge of data silos. CIOs are increasing budgets for data infrastructure, aiming for notable efficiency improvements in AI project timelines with integrated platforms like Salesforce Agentforce.
A recent survey conducted by ZDNet has revealed that a significant majority of Chief Information Officers (CIOs), specifically 93% of respondents, are planning to implement AI agents within their organisations over the next two years. With 1,050 CIOs participating in the survey, the data reflects a strong commitment among IT leaders to eliminate data silos as part of their strategy to adopt this emerging technology.
The survey highlights that participants reported an average of 897 applications in use, with 45% of respondents indicating they utilise over 1,000 applications. This multitude of applications complicates IT teams’ efforts to create a seamless user experience. As it stands, only 29% of enterprise applications have the capability to interact and share information across the business, underscoring the urgency for improved integration.
To facilitate the anticipated rise in AI adoption, enterprise CIOs are devoting a considerable 20% of their budgets to enhancing data infrastructure and management. This allocation is notably four times greater than their spending on AI technology, which sits at 5%.
Defining AI agents, ARK Invest notes that these technologies are expected to transform digital application adoption and human-computer interactions by understanding user intent through natural language, planning with reasoning and context, executing tasks through appropriate tools, and improving through iteration and continuous learning. Forecasts from ARK project that software deployment per knowledge worker will grow significantly, estimating that global software expenditure could rise from an average annual growth rate of 14% over the past decade to between 18% and 48%.
For businesses aiming to maximise their return on investment from agentic AI technologies, the technology research firm Valoir outlines that these systems can deliver significant advantages by automating intricate tasks and interactions without human oversight. However, the development of such complex AI systems presents challenges. Valoir’s research indicates that leveraging a platform specifically optimised for agentic AI development, like Salesforce Agentforce, can enable businesses to create autonomous AI agents approximately 16 times faster than conventional methods, while simultaneously improving accuracy rates by 75%.
The research delineates seven distinct phases of agentic development, which encompass model setup, data and application integration, prompt engineering, AI guardrails and security implementations, user interface and workflow/application development, tuning, and ensuring data accuracy. A notable finding from Valoir’s analysis is the contrast between a Do-It-Yourself (DIY) approach and the use of a fully integrated platform equipped with embedded agentic AI capabilities. The study shows that organisations pursuing a DIY methodology typically rely on pre-built models that necessitate a setup time ranging from three to twelve months, compared to the significantly reduced setup time achieved through platforms like Agentforce, which are pre-integrated and require no more than an average of 7.5 times less setup time.
Additionally, Valoir noted that organisations opting for open-source solutions often spent at least a month deciding on a Retrieval-Augmented Generation (RAG) approach before engaging in the integration of document handling and generative models, extending the overall processing time to several additional months. Conversely, data and application integration using Agentforce was completed in a matter of weeks, averaging approximately three and a half times faster than other approaches.
Trust, a crucial element in the transition from generative to agentic AI applications, was highlighted as a significant aspect where those with extensive development and data science expertise required over a year to develop a trust layer. The study also indicated that accuracy in data is a pivotal factor influencing the time required to establish and train AI systems to achieve satisfactory levels of correct responses. For simpler tasks, accuracy rates were identified at 50% for DIY methods compared to 95% for solutions employing Agentforce. In more complex tasks, like sales coaching, these rates were 40% for DIY versus 95% for Agentforce.
In summary, findings from Valoir underscore that the average duration of DIY projects spans 75.5 months, while implementations using the Agentforce platform can achieve productive accuracy in as little as 4.8 months—indicating a substantial efficiency improvement in AI project timelines.
Source: Noah Wire Services
Subscribe to Industry Updates
