Companies eager to adopt AI face data quality challenges

New research from Semarchy has found a significant gap between companies' aspirations to invest in artificial intelligence (AI) and the reality of their data quality concerns.

According to the research, 74% of businesses plan on investing in AI initiatives this year. However, only 46% of respondents trust the quality of the data that will feed these AI systems. The study surveyed 1,050 senior business leaders across the United States, United Kingdom, and France.

This discord underscores a major challenge for businesses eager to implement AI without fully ensuring data readiness. A vast majority, 98%, confessed to encountering data quality issues related to AI. Predominantly cited problems include data privacy and compliance issues (27%), duplicate records (25%), and inefficient data integration (21%).

The repercussions of these data quality issues have been significant. They have led to diminished trust in AI outputs for 19% of businesses, presenting crucial risks to AI's credibility and widespread adoption. Some firms have also reported delays in rolling out new projects (22%) and rising costs (20%).

The research highlights the disconnect between executive ambitions for AI and the practical execution of these goals. Less than half of business leaders (46%) feel confident in achieving their AI objectives this year, with this figure dropping to 35% for Chief Data Officers (CDOs). This cautious outlook is likely attributed to CDOs' deep understanding of the data hurdles that need to be overcome for AI to provide real value.

Craig Gravina, Chief Technology Officer at Semarchy, commented, "The enthusiasm for AI adoption from businesses is clear, but our findings reveal a troubling gap between ambition and reality. Businesses are eager to embrace AI, yet many lack the data quality necessary for success. Deploying AI at scale on a bad foundation will only magnify business risks and lead to wasted investments."

Gravina further added, "The next steps here are logical: look at the business case for AI closely and assess the readiness and risk of their data before jumping in headfirst. We need to reshape the way businesses view enterprise data and AI ecosystems in order to fuel AI-driven innovation while ensuring transparency, security, and governance."

Gilles Corcos, CIO Sales & Marketing at Elis, shared experiences from his company's strategy: "By strategically combining the capabilities of Artificial Intelligence (AI) and Semarchy Master Data Management (MDM), we have established a self-reinforcing cycle of data quality improvement. This synergy allows our data stewards to redirect their focus from time-consuming, repetitive tasks towards more strategic, high-value activities. This not only enhances the overall quality and reliability of our data but also translates to substantial time and cost savings for Elis."

Jean-Yves Falque, Founder and Executive Chairman at Apgar, also remarked on the necessity of strong data foundations: "AI is a social and industrial revolution in motion, and every company must embrace it to stay competitive. Yet, too many initiatives fail due to a lack of digitalization and insufficient governance and data quality management. Laying a strong foundation is essential to building trust and ensuring an effective, responsible rollout. Unlike BI, AI will democratize data usage and bring data quality and governance to the core of business operations. Like any transformation, it requires education and commitment, empowering every 'data citizen' to contribute to making it a true value driver—essential for the long-term success of their company."

The study also indicates some ambiguity surrounding corporate AI leadership and governance. While 37% of CIOs, 30% of CTOs, and 23% of CEOs see themselves as chiefly responsible for AI strategy, only 15% of CDOs view themselves in this role. Notably, just 6% of CEOs believe CDOs should be in charge of AI initiatives, which is a particularly low figure given AI's heavy reliance on data quality and governance.

Craig Gravina emphasized the benefits of a collaborative approach by saying, "AI leadership shouldn't be driven by siloed thinking or short-term priorities. Instead, it should be based on a range of expertise and experience and, of course, high-quality data. The research shows that just 7% of organizations have a cross-functional team driving AI strategy. A collaborative approach is essential to align technical capabilities with business objectives, but strong data leadership remains critical."

The survey also touched on ethical, regulatory, and security concerns as businesses seek to accelerate AI adoption this year. Less than half of the firms (45%) are actively working on mitigating AI bias, and a similar proportion (47%) admitted employees use non-private AI environments, which could involve handling sensitive company data.