In collaboration withDatabricks
Four years represent an eternity in the realm of artificial intelligence. Since this study’s initial release in 2021, the capabilities of AI have been evolving rapidly, and this evolution has not decelerated following the generative AI breakthrough. For instance, multimodality— the capacity to comprehend information not just as text but also as audio, video, and various unstructured formats—is increasingly becoming a standard feature in AI models. The ability of AI to reason and operate independently has also expanded, with organizations now beginning to engage with AI agents capable of performing such tasks.
In the midst of these changes, one constant persists: the quality of an AI model’s outputs is intrinsically linked to the data
that nourishes it. Data management technologies and methodologies have advanced as well, yet this study’s second edition indicates that the majority of organizations are not utilizing those advancements quickly enough to match AI’s progression. Consequently, due to this and other obstacles, a relatively small number of organizations are achieving the desired business outcomes from their AI strategies. No more than 2% of senior executives we surveyed believe their organizations are highly effective in terms of delivering results from AI.
To assess the degree to which organizational data performance has enhanced as generative AI and other advancements have emerged, MIT Technology Review Insights surveyed 800 senior executives in data and technology. We also carried out extensive interviews with 15 technology and business leaders.
Key insights derived from the report include:
• A small number of data teams are keeping up with AI. Organizations are not performing any better today in executing data strategy compared to the pre-generative AI era. Among those polled in 2025, 12% categorize themselves as data “high achievers” compared to 13% in 2021. While a scarcity of skilled professionals remains a barrier, teams also face challenges accessing current data, tracking lineage, and managing security complexities—critical factors for AI success.
• Consequently, AI is not yet operating at full capacity. There are even fewer “high achievers” in the realm of AI. Only 2% of respondents assess their organizations’ AI performance as high today concerning tangible business outcomes. Many are still grappling with scaling generative AI. While two-thirds have implemented it, just 7% have done so extensively.
This content was generated by Insights, the custom content division of MIT Technology Review. It was not authored by the editorial staff of MIT Technology Review. It underwent research, design, and writing by human writers, editors, analysts, and illustrators. Any AI tools utilized were restricted to secondary production processes that were thoroughly reviewed by humans.
