Artificial intelligence has advanced rapidly since 2021, but most organizations are still failing to turn those gains into measurable business results, according to a new study by MIT Technology Review Insights.
The report, based on a survey of 800 senior data and technology executives and interviews with 15 business and technology leaders, finds that while AI tools—particularly generative and multimodal models—have become more capable, the underlying data foundations needed to support them have not kept pace.
“Four years is a lifetime in AI,” the report notes, pointing to breakthroughs in multimodality and the emergence of autonomous AI agents that can reason and act with limited human input. Yet fewer than 2 percent of senior executives rate their organizations as highly effective at delivering business value from AI.
Data shortcomings remain the central bottleneck, according to the study. Just 12 percent of organizations describe themselves as data “high achievers” in 2025, virtually unchanged from 13 percent in 2021, despite heavy investment in AI technologies. Persistent challenges include shortages of skilled data talent, difficulty accessing fresh and trusted data, weak data lineage, and growing security complexity.
As a result, AI adoption has been shallow rather than transformative. While nearly two-thirds of organizations have deployed generative AI in some form, only 7 percent say it has been rolled out widely across the business. Performance drops further when executives assess AI’s ability to deliver clear, measurable outcomes.
The study also highlights weak alignment between data and AI strategies. Fewer than half of respondents—46 percent—express confidence that their data strategies can adapt to future AI advances. Many organizations operate separate governance models for data and AI, and lack unified technology platforms, limiting scalability.
The rise of so-called agentic AI—systems capable of taking autonomous actions—raises the stakes further. MIT Technology Review Insights warns that these models demand stronger data governance, explainability, and security. Most organizations are moving cautiously, focusing on building reliable data foundations before deploying such systems at scale.
At the same time, AI is beginning to reshape data management itself. About 67 percent of surveyed organizations already use AI-powered data tools, with the remainder planning adoption within two years. Meanwhile, 69 percent report using data intelligence tools to map assets, break down silos, and strengthen governance.
The report concludes that while AI adoption is now widespread, the gap between ambition and execution remains wide. Without faster progress on data quality, governance, and strategy alignment, many organizations risk falling further behind as AI capabilities continue to advance.
