As AI and advanced analytics technologies rapidly spread, the standards for corporate competition are changing. The critical question is no longer whether to adopt AI, but whether a company has the environment to create real business value based on reliable data. In other words, the key to corporate competitiveness in the AI era lies in 'Data Readiness.'
Many companies recognize the importance of data and are accelerating their investments in AI, but the reality is more challenging than expected. A recent survey by Cloudera, which included over 1,200 IT leaders across 14 countries, found that 89% of global respondents (93% in South Korea) believe that executives view data infrastructure as a core issue for AI expansion.
Additionally, 86% reported having established a data strategy linked to business objectives, and the same percentage of companies indicated they are increasing investments in data infrastructure and cloud services. This indicates that data is no longer just an asset of the IT department but has become a central agenda for corporate management.
However, investment and interest do not automatically translate into data competitiveness. The biggest obstacles are data silos within organizations and low data accessibility. The survey revealed that only 30% of companies reported having fully integrated data, while more than half indicated that their data is only partially connected.
In South Korea, the situation is even more severe, with 80% of responding companies stating that their data is not fully connected. Most data is dispersed across different systems, making it difficult to utilize it company-wide.
This lack of data readiness leads to poor performance. According to the report 'Overcoming the Complexity of AI Data Readiness' by Cloudera and Harvard Business Review Analytics Services, 73% of business leaders struggle with data preparation for AI, citing data silos and difficulties in source integration (56%), lack of data strategy (44%), quality and bias issues (41%), and regulatory constraints (34%) as the main causes.
In fact, a study by the RAND Corporation found that about 80% of AI projects fail to achieve the expected business outcomes, nearly double the failure rate of typical IT projects.
International examples illustrate how detrimental a lack of data readiness can be. The U.S. real estate platform Zillow ambitiously pursued its 'Zestimate' AI model to predict home prices for its Zillow Offers business. However, the model failed to timely reflect rapidly changing market data post-pandemic, leading to continued overvaluation of home prices. Ultimately, in the third quarter of 2021 alone, Zillow incurred a loss of $421 million and had to shut down the business, reducing its workforce by 25%. This case demonstrates that even the most sophisticated algorithms can become detrimental without fresh and reliable data.
Conversely, there are success stories from companies that invested in data integration first. Global investment bank Standard Chartered built an enterprise data fabric to efficiently manage dispersed data and varying regulatory environments across countries. They integrated management from data collection to access rights, security policies, and data lineage, establishing a foundation for analyzing and utilizing enterprise data for AI while complying with national data sovereignty and financial regulations. Their standardized data pipeline reduced the data pipeline development cycle by about 40%, significantly enhancing data utilization speed and operational efficiency.
However, the challenges do not end there. As companies rush to adopt AI, the phenomenon of 'AI silos'—where different departments invest in overlapping tools—has emerged as a new issue. When tools vary by department, the consistency of data and models is compromised, complicating governance. Experts advise that organizations need to build an integrated data and AI platform to address this issue.
AI can only deliver results when it is based on high-quality data. If data is scattered across multiple systems or if access rights and governance are inadequate, even the most advanced AI models will struggle to achieve the desired outcomes. Ultimately, the essence of data readiness lies not in how much data is collected, but in how well it is integrated, managed in a trustworthy manner, and safely utilized when needed.
The competition in AI is now a competition for data, not just models. Companies must not only store data but also connect it as a single asset, enabling free utilization in AI and analytics environments, regardless of its location.
The divergence between Zillow's failure and Standard Chartered's success ultimately comes down to one factor: how effectively they managed data as a trustworthy asset. Only companies that build an 'AI-Ready Data Platform' will be able to create new business value and secure a sustainable competitive advantage in a rapidly changing market.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.