With the current landscape of innovation in AI, data has never been more crucial for enterprise. Consumption data can be used to train machine learning models that forecast revenues, while documentation about products can be used in a chatbot to assist customers or sales teams with up-to-date, accessible information. At the cutting edge, AI agents given a task can rapidly build on large-scale data sources, extracting insights and hidden patterns from them to improve the way a business works, while ‘embodied AI’, with digital twins and robotics, interfaces with data to bring the digital world to the physical one.
However, for a business to actually reap the benefits of AI, the data foundations underneath it must be optimised. If they are not ‘AI-ready’, AI implementations risk negative outcomes that can undermine investment and harm business activity.
In this research, we sought to build on our previous work and define what AI readiness means in the context of enterprise data. We focused on creating specific criteria for organisations to follow, providing 21 actionable recommendations for the design of datasets, metadata, infrastructure, and governance to be optimised for AI.
To build this framework, we utilised insights from a comparative analysis of 23 publications, a set of expert interviews with 8 participants, and the lived experiences of the ODI team which has worked on data practices for over a decade. We present the requirements alongside contextual information in the main report and summarise them in a visual framework that can be used to assess and improve data architectures for AI.
This research was conducted as a component of our partnership with SAP on the Interchange for Data and Enterprise AI (IDEA). As with our previous framework for AI-ready data, it will be refined and updated over time, and it will also be used as a basis for further research in which we dive deep with specific enterprises to explore their AI and data ecosystems. If you would like to be involved in these next steps, please let us know by emailing [email protected].
