Clouderaspecialized in promoting AI through data regardless of the environment, ha Published Los Results from Su Studio The Evolution of Ai: The Stated of Enterprise Ai and Data Architectureconducted from a survey among more than 1,500 IT leaders, and that examines how AI is being adopted in all types of organizations, what data architectures they use and the new challenges that have emerged in 2025.
According to the study, companies recognize that they are still advancing towards obtaining the greatest potential and roi in AI. With respect to 2024, 24% of respondents ensure that their company has a culture completely driven by the data, compared to 17% of 2024.
Despite this increase, accessibility is still an obstacle: only 9% of organizations ensure that all their data are prepared and accessible to AI initiatives. Another 38% ensure that most are. The main technical limitations that identify for this include data integration (37%), storage performance (17%) and computer power (17%).
96% of those responsible assured that AI is integrated, at least to some extent, in their basic company processes. This is an increase with respect to 2024, when this figure was 88%. In addition, 70% of respondents said they had achieved remarkable success with their AI initiatives, and only 1% have not seen results yet.
The study also analyzes what types of AI are being used. The most used, in 60%of cases, is the generative AI, followed by the Deep Learning (53%) and the predictive (50%). As for the confidence in these models, 67% of those responsible for you feel more prepared than in 2024 to manage new forms of AI, such as agricultural AI.
The hybrid approach to data architecture has become a standard, since it allows companies to the flexibility they need to manage AI in local and cloud environments. For 62% of respondents, the main advantage of this system is safety, while for 55% is the improvement of the management DEW data, and for 54% the improvement of the analytics.
As for data storage, 63% keep them in the private cloud, 52% in the public, and 42% in a Warehouse data. And regarding the integration of AI, continue to worry its safety. 50% of respondents said that data leakage during models training is an important concern, 48% mentioned as a problem unauthorized access to data and 43% The use of third -party AI tools that are not safe.
However, 24% of respondents said they have a lot of confidence in the ability of their organization to protect the data used in AI systems, 53% showed a lot of confidence, and 19%, quite confidence.
Sergio Gago, CTO Global of Clouderastressed that «in Only one year, AI has gone from being a strategic priority to become an urgent need, redefining the rules of competition. However, our study reveals that companies continue to face great challenges in terms of safety, regulatory compliance and data exploitation. Many of them are stagnant in the concept test phase. The clouding mission is to bring AI to the data wherever they are (public cloud, private and local cloud), while guaranteeing a total lineage and confidence. We offer companies the total control and trust necessary to extract information from 100 % of their data«.
