Gartner identifies 'semantic foundation' as key condition to boost AI agent performance.
Published2026.05.12 12:01.jpg)
Gartner stated in Seoul on May 12 that if semantics are overlooked in the operation of AI agents, accuracy and efficiency will decrease, and unnecessary costs and data and AI governance risks may increase.
AI agents make judgments and generate results based on information input at each stage of the workflow. If they fail to understand the meaning of the data and its relationships with one another during this process, there is a high likelihood that the resulting answer will differ from the actual business context.
Gartner advised that a 'context layer' should be built into the data and analytics infrastructure to address this. The explanation is that existing schema-centric data models alone make it difficult to sufficiently convey the business context and data meaning required by agentic AI.
Rita Salam, a senior VP analyst at Gartner, explained that the performance of agentic AI depends on the context, including the semantic representation of the data. She noted that if the relationships and rules of organizational data are not clearly understood, there is a greater likelihood that AI agents will be hallucinating or produce biased results.
Gartner projected that by 2027, organizations that prioritize semantics on AI-ready data will be able to increase the accuracy of agentic AI by up to 80% and reduce costs by up to 60%.
The importance of semantic transparency is expected to grow in the future regulatory environment as well. Gartner predicted that boards of directors will come to view semantic governance not merely as a technology management issue, but as a strategic risk and a competitive opportunity.
As the use of AI agents becomes more widespread, the ability to consistently define and manage the meaning of data is expected to become a foundational factor determining a company's AI operational performance.
관련뉴스
Gartner: “Global IT spending to grow 10.8% by 2026… Continued expansion of AI infrastructure investment”
Gartner forecasted that global IT spending in 2026 will reach $6.15 trillion, representing a 10.8% increase from the previous year. Continued AI infrastructure investment is expected to sustain double-digit growth rates in data center systems and server spending. Software spending growth forecasts were slightly revised downward, but generative AI model spending is projected to maintain growth exceeding 80%. In contrast, device spending growth is expected to decelerate due to dampened demand resulting from rising memory prices.
2026.02.09by 배종인 기자
.jpg)
Nozomi Networks Named a Leader in Gartner's CPS Protection Platforms for the Second Consecutive Year
Global IT research firm Gartner has recognized Nozomi Networks as a Leader in its "2026 Cyber-Physical System Protection Platform Magic Quadrant" report, reaffirming Nozomi Networks' position in the industrial control systems and cyber-physical systems (CPS) security market.
2026.03.10by 배종인 기자
AI Changes Hiring Standards… Gartner Forecasts Reorganization of Data and Analytics Organizations
Gartner predicts that the widespread adoption of AI will transform how enterprises operate their data and analytics organizations across hiring, productivity tools, governance, and leadership. By 2027, 75% of hiring processes will evaluate applicants' AI proficiency, and generative AI and AI agents are projected to drive an $58 billion shift in the productivity software market. In the long term, AI data generated from physical environments will surge significantly, and autonomous AI agent-based governance automation and the establishment of a universal semantic layer are expected to emerge as critical competitive imperatives for enterprises.
2026.03.16by 명세환 기자
Gartner: “Corporate Burden Continues Even as AI Token Prices Drop”… Proliferation of Agents Changes Cost Structures
Gartner forecasts that the inference unit cost of ultra-large language models (LLMs) will decline by over 90% by 2030 compared to 2025. Improvements in semiconductor performance and enhanced model design efficiency are cited as key drivers. However, the analyst also noted that enterprise AI cost burdens may not decrease proportionally. This is because as more complex functions such as AI agents proliferate, the number of tokens required per task will increase significantly, potentially raising overall inference costs. Consequently, future enterprise competitiveness is expected to depend not on reliance on a single large model, but rather on the capability of "multi-model orchestration"—strategically leveraging smaller models and high-performance models according to specific business requirements.
2026.03.30by 배종인 기자
Gartner: Global PC Shipments Up 4% in Q1... Impact of Expanded Inventory
According to Gartner, global PC shipments in the first quarter of 2026 reached 62.8 million units, representing a 4% increase year-over-year. However, this growth was largely driven by inventory expansion in anticipation of rising memory prices and potential increases in component costs, rather than actual demand expansion. Lenovo, HP, Dell, and Apple maintained their positions as the top four vendors, while ASUS surpassed Acer to claim the fifth position. Apple recorded the highest growth rate among major vendors, driven by MacBook Neo demand.
2026.04.13by 명세환 기자
