
IT leaders face data challenges with generative AI, such as data quality, usability, access, and pre-existing data challenges in machine learning projects1. Data often exists in different silos and incompatible formats, requiring preprocessing, cleaning, anonymizing, and consolidating before use in machine learning. Data engineering and ownership management remain important challenges for most machine learning and AI projects.

The AI survey by Dataiku and Cognizant measures the current state of AI implementation, budget allocation for generative AI initiatives, infrastructure barriers, regulatory compliance challenges, operational costs, data infrastructure challenges, and the level of AI maturity within organizations.