Affordance–Experimentation–Actualization- Evolution for AI-based Support of Innovation

Artificial Intelligence (AI) holds significant promise for supporting decision-making in innovation ecosystems. However, this potential is constrained by fragmented and limited data, often scattered across industrial actors and restricted by concerns of sovereignty, confidentiality, and regulation. These limitations hinder the development of robust and generalizable AI system to evaluate early-stage ventures. To address these challenges, we introduce datA-Inno, an adaptive AI-based decision-support system designed to assist decision-makers in evaluating early-stage ventures operating within fragmented data environments. To formalize the operational logic of datA-Inno, we extend the traditional Affordance–Experimentation–Actualization (AEA) framework with an
explicit Evolution phase, resulting in the 2*AE model. This extension enables improved management of AI system transitions across multiple contexts without premature appropriation. We illustrate the applicability of datA-Inno and the 2*AE
framework through scenarios in which the AI system evolves
from prompt-based models to federated and decentralized ar-
chitectures, demonstrating how it can support effective venture
evaluation.