Project

Neotec Shadow Experiment

Using AI and experimentation to improve public funding decisions with CDTI

Science and Innovation Funding Experimental organisations

Artificial intelligence (AI) is rapidly transforming how governments and public organisations operate. Research and innovation funding agencies are increasingly exploring how AI can support activities such as proposal assessment, reviewer assistance and funding allocation. Yet, despite growing interest, there is still limited evidence on whether these technologies actually improve decision-making, how assessors interact with them, and how to introduce them responsibly into high-stakes public funding processes.

To address these questions, IGL partnered with CDTI, Spain’s national innovation agency, and a group of researchers to design and implement a large-scale shadow experiment within Neotec, one of Spain’s flagship programmes supporting early-stage technology companies.

Rather than introducing AI directly into the official funding process, the project created a parallel experimental environment where alternative human-AI evaluation models could be tested without affecting funding decisions. This allowed the team to generate evidence under realistic operating conditions while preserving the fairness and integrity of the programme.

Project activities

The project combines policy experimentation, artificial intelligence and behavioural research to explore how AI can support evaluators in assessing the social impact of innovation projects.

Key activities include:

Designing a shadow experiment to evaluate AI-assisted proposal assessment without influencing real funding decisions.
Developing a dedicated AI system capable of generating structured social impact assessments for Neotec applications.
Comparing the predictive performance of different evaluator groups and AI-supported assessments using real funding proposals.
Analysing how evaluators respond to AI-generated recommendations and whether these tools influence decision-making.
Generating practical lessons on how public funding organisations can experiment with emerging technologies before integrating them into operational processes.

Expected impact

The project aims to contribute to a growing international discussion on the future of public funding organisations in the age of AI.

Beyond understanding the performance of AI-assisted evaluation, the project demonstrates how experimentation can help governments adopt emerging technologies in a safe, evidence-based and responsible way. By developing and testing a shadow experiment within a real funding programme, the project provides a practical model that other funding agencies can adapt to evaluate new technologies without compromising fairness, legitimacy or public trust.

More broadly, the project contributes to IGL’s mission of helping governments generate better evidence, build experimentation capacity and improve public policy through rigorous testing and learning before large-scale implementation.