Less than a quarter (22%) of organizations believe their artificial intelligence deployments are strategic in nature, according to the latest edition of the Anaconda State of Data Science & AI survey. The comprehensive report explores how data scientists, developers, and engineers around the world are using open-source tools, where value is measured, and what obstacles still hamper enterprise AI adoption.
The survey reveals that productivity improvements (58%) and cost savings (47%) are the dominant metrics used to gauge AI return on investment (ROI). Far fewer—just 25%—focus on measures such as go-/no-go decisions, error reduction or risk mitigation, and 13.5% of respondents admitted their organizations do not measure ROI at all or are unsure how they do. An unclear or absent strategy, the report warns, may be limiting how productive AI initiatives become.
When it comes to scaling AI into production, respondents cited several persistent barriers. Data-quality and pipeline consistency topped the list at 45%; 40% said security, privacy and regulatory compliance were the greatest challenge; 39% pointed to scaling inference and compute-cost management; and 33% flagged cross-team collaboration and communication gaps. These structural issues, Anaconda argues, reflect misalignment between teams, technologies and strategic goals that slows AI progress.
On the open-source front, signs of progress are strong. Some 92% of respondents say they already use open-source AI tools or models, and 76% claim open-source holds higher priority this year than the last. Anaconda argues that “open innovation” is helping organizations build a stronger foundation for AI, even if strategic uptake lags.
The findings signal a new phase in enterprise AI: one in which momentum is high but depth remains shallow. As companies move out of experimentation into deployment, the report suggests they must shift from tactical use cases to purpose-driven strategies, align governance and infrastructure, and tie AI more closely to real business outcomes.
