Skip to content 
AI Project Guide
The problem we address
While artificial intelligence offers enormous potential to improve public services and policies, many public sector projects fail or create risks due to a lack of clarity on where to start, how to make informed decisions and how to manage ethical, legal, and operational risks. There is often a gap between abstract principles and their concrete application in real projects.
What do we do?
We developed a practical guide for implementing data science and AI projects in the public sector, which translates ethical principles and regulatory frameworks into operational, measurable, and auditable decisions. The guide accompanies institutions throughout the entire project cycle, from problem definition to the management of risks and responsibilities.
How do we do it?
The guide is structured into a three-phase, ten-step roadmap, with guiding questions, decision gates, and minimum deliverables at each stage. It allows for evaluating whether AI is the right solution compared to other alternatives, defining technical and equity metrics, establishing usage limits and safeguards, and documenting the traceability of decisions. Additionally, it creates a common language across public policy, data, IT, legal, public procurement, and communication areas, strengthening project governance.
The problem we address
While artificial intelligence offers enormous potential to improve public services and policies, many public sector projects fail or create risks due to a lack of clarity on where to start, how to make informed decisions and how to manage ethical, legal, and operational risks. There is often a gap between abstract principles and their concrete application in real projects
What do we do?
We developed a practical guide for implementing data science and AI projects in the public sector, which translates ethical principles and regulatory frameworks into operational, measurable, and auditable decisions. The guide accompanies institutions throughout the entire project cycle, from problem definition to the management of risks and responsibilities.
How do we do it?
The guide is structured into a three-phase, ten-step roadmap, with guiding questions, decision gates, and minimum deliverables at each stage. It allows for evaluating whether AI is the right solution compared to other alternatives, defining technical and equity metrics, establishing usage limits and safeguards, and documenting the traceability of decisions. Additionally, it creates a common language across public policy, data, IT, legal, public procurement, and communication areas, strengthening project governance.
