Dealio Research Project

Harmony

AI-Driven Hedging Intelligence with Human Expert Validation

A research and innovation project developing an explainable, human-centred AI framework for financial risk management built to align with the EU AI Act and modern market realities.

Project Number: NPD-CAPBLD/0825/0003

About the project

HARMONY is a Dealio-led research and innovation project developing an AI-driven hedging intelligence framework for financial risk management. The project moves beyond traditional risk models that rely on static assumptions and historical data, building instead a system that combines domain-specific machine learning with contextual financial, macroeconomic, and geopolitical signals. A core design principle of HARMONY is integrating human expertise into the AI workflow, ensuring algorithmic outputs are interpretable, trustworthy, and aligned with regulatory frameworks such as the EU AI Act.

HARMONY is a Dealio-led research and innovation project developing an AI-driven hedging intelligence framework for financial risk management. The project moves beyond traditional risk models that rely on static assumptions and historical data, building instead a system that combines domain-specific machine learning with contextual financial, macroeconomic, and geopolitical signals. A core design principle of HARMONY is integrating human expertise into the AI workflow, ensuring algorithmic outputs are interpretable, trustworthy, and aligned with regulatory frameworks such as the EU AI Act.

Main objectives

  • Design and develop an AI-driven hedging intelligence framework tailored to modern financial markets.

  • Integrate machine learning models with contextual feature engineering across financial, macroeconomic, and geopolitical signals.

  • Embed a human-in-the-loop validation layer so financial experts review AI-generated outputs before action is taken.

  • Ensure interpretability, transparency, and alignment with emerging regulatory frameworks, including the EU AI Act.

  • Validate the framework in realistic conditions and prepare it for operational deployment.

Expected results

  • A validated hybrid intelligence framework combining algorithmic analysis with expert oversight.

  • Context-aware hedging recommendations and anomaly detection signals for faster, more informed decisions.

  • An explainable and auditable AI system suitable for regulated financial environments.

  • A reusable methodological blueprint for human-centred AI in financial risk management.

  • Strengthened capacity within Dealio to deliver next-generation, trustworthy AI products.

Acknowledgement

The project (NPD-CAPBLD/0825/0003) was funded by the Research and Innovation Foundation, under the «ENTERPRISES CAPACITY BUILDING IN NEW PRODUCT DEVELOPMENT Programme» and through the Recovery and Resilience Facility of the NextGenerationEU instrument.