Cedar has been featured in the latest HiPEAC magazine (June 2026 Data & AI special issue).
Here the complete article "Transforming EU governance The CEDAR project and the dawn of a data-driven public procurement era"
---------------------------------
Public procurement is the engine room of the European economy, accounting for approximately 14% of the European Union’s gross domestic product (GDP) – over €2 trillion annually. Yet for decades this vast expenditure has been managed through fragmented, localized systems that often lack transparency, making them vulnerable to inefficiency, fraud, and corruption. Enter the CEDAR project (Common European Data Spaces and Robust AI for Transparent Public Governance), a Horizon Europe initiative (2024–2026) designed to revolutionize the full lifecycle of procurement data management. By leveraging advanced artificial intelligence (AI) and the transformative power of Common European Data Spaces (CEDS), CEDAR is going beyond digitizing records to build a cross-border intelligence layer that ensures that public money is spent accountably and effectively.
From fragmented data to FAIR intelligence
The primary hurdle in public governance has never been a lack of data, but rather its ‘siloed’ nature. Procurement notices, contract awards, and invoices are often scattered across thousands of regional portals in varying formats, from structured databases to unsearchable PDFs. CEDAR addresses this through a robust data management cycle rooted in the FAIR principles (findable, accessible, interoperable, and reusable). The project implements a sophisticated data operations (dataOps) and machine learning operations (MLOps) pipeline that automates the ingestion, cleaning, and harmonization of data from diverse sources (text and media). Using a focused ontology and a knowledge graph that links together tenders, bids, and actors, CEDAR creates a ‘common language’ for procurement, ensuring that a contract award in Italy can be semantically compared to one in Slovenia or Ukraine.
Since different organizations are at different levels of maturity in terms of digitization, this cycle doesn't start with digitized records and it doesn’t end with data storage. CEDAR provides tools and services for digitizing scanned documents using AI, processing and cleaning the data to be fitted to the knowledge graph, and data stored in appropriate pseudo-anonymized repositories. Finally, through secure connectors and application programming interfaces (APIs), the project ensures that highquality, pseudonymized datasets are ready for the next crucial stage, i.e. deep analysis via AI to detect misuse indicators and further investigation alerts. CEDAR here is at the heart of the CEDS mission to produce and analyse data that improve intelligence analysis.
Advanced AI for continuous public integrity checks
At the heart of the CEDAR architecture lies a suite of advanced AI models designed to act as a sentinel. Unlike traditional oversight, which often relies on manual audits of a tiny fraction of contracts, CEDAR’s AI analyses 100% of the available data in real time. The project utilizes explainable AI (XAI) and graph-based data modelling to detect patterns that are invisible to the human eye. The system operates on two levels of triggers:
- Simple indicators These are the ‘red flags’ often associated with compliance. They include indicators like ‘single-bidding’ (where only one company applies for a tender), ‘purchase splitting’ (breaking a large contract into smaller ones to avoid oversight thresholds), or ‘negotiated procedures’ without prior publication.
- Complex indicators These are where CEDAR’s AI truly shines. By integrating external data sources – such as company registries and historical bidding behaviour – the system identifies hidden ‘collusive networks’. It can detect bid rotation (where companies take turns winning), price manipulation, and tailored tenders (where requirements are so specific they can only be met by one favoured bidder).
When these indicators pass a certain threshold of ‘normality’, they trigger automated alerts for further investigation by public authorities, shifting the focus from random sampling to evidence-based risk management.
The central vision: Common European Data Spaces (CEDS)
While AI provides the brainpower, the CEDS provide the infrastructure for collective intelligence. This is the central concept of the CEDAR vision: the project may serve as a bridge to the Public Procurement Data Space (PPDS) and other sectoral spaces like the European Health Data Space and the Green Deal Data Space. In the past, an anti-corruption tool developed in one Member State was only as good as the local data it was trained on. CEDAR breaks this limitation. By connecting to the CEDS, the project enables a trans-European feedback loop:
- Sharing aggregated results: instead of sharing raw, sensitive data, Member States can share the ‘aggregated results’ of their analyses. If a specific pattern of fraud is detected in the medical procurement sector in Italy, the parameters of that ‘indicator’ can be shared across the CEDS.
- Updating EU-wide indicators: other nations can then update their local AI models with these new parameters. This creates a dynamic, evolving defence system where an attempt at corruption in one corner of the EU instantly makes the entire union more resilient.
- Cross-sectoral enrichment: by linking procurement data with the European Health Data Space (EHDS) or the Green Deal Data Space, CEDAR can monitor, for example, whether ‘green procurement’ targets are actually being met or if pharmaceutical spending is optimized across borders.
Progress towards greater transparency
Currently in its last year, the CEDAR project – a consortium of 31 partners including research institutions, tech leaders, and non-governmental organizations (NGOs) – is validating its technologies through high-impact pilots. Pilots in Italy, Slovenia and Ukraine focus on the health sector, analysing historical bidding data from hospitals and ministries to identify risk patterns in medical supply chains. The project’s progress is measured not just in lines of code, but in the creation of interoperable connectors and AI tools that allow any public organization to plug into the CEDS ecosystem and get deep analysis. By 2026, CEDAR aims to have successfully shared novel AI-driven indicators, high-quality procurement datasets and enriched different CEDS, providing a blueprint for the EU’s ‘Digital Decade 2030’.
Conclusion
The CEDAR project represents a fundamental shift in how we view the relationship between the state and data. By moving from reactive oversight to proactive, AI-driven management, the EU is setting a global standard for transparent governance. Through the use of CEDS, CEDAR ensures that public procurement is no longer a collection of isolated transactions, but a unified, intelligent ecosystem. In this new era, data can become a tool for building a more fair, competitive, and trustworthy European future.
--------------