This framework provides a methodology for characterizing and classifying any Generative AI–based solution across nine dimensions. The goal is to create a single vector that classifies each specific solution, considering its architecture, training model, how it is used by the organization, the type of data and information it handles, among other dimensions, in order to measure the operational risk of that solution within an organization.

The assessment of risk depends both on dynamic variables and on the intrinsic qualities of the AI model, and the result is aligned with the specific requirements defined in the organization’s policies for risk management and analysis. The purpose is to enable the organization to characterize the risk involved in adopting any Generative AI–based solution, whether developed internally or acquired from the market, thus allowing the organization to take responsibility for the decision to use or not use that solution.