The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

AI Copilot for the National Post Provider


Added by:   OECD analyst
Added on:   23 Sep 2026
Updated by:   OECD analyst
Updated on:   23 Sep 2026

Posta Kopilot is an AI-powered mobile application that assists postal workers with manual mail sorting at the National Post of Serbia (Pošta Srbije), which processes over 2.1 million pieces daily. By interpreting address information and guiding sorting decisions in real time, it has increased sorting efficiency by over 20% and cut new-employee onboarding from several months to a few weeks. Developed through Serbia's Innovation Fund GovTech programme with LambdaWorks.

Name in original language

Pošta Kopilot

Initiative overview

Posta Kopilot is an AI-powered mobile application designed to enhance the efficiency of manual mail sorting at the National Post of Serbia (Pošta Srbije). Serbia's national postal service handles over 2.1 million pieces of mail daily, and a significant portion of the sorting process remains manual. Sorting mail traditionally requires extensive knowledge of postal codes, delivery routes, and address structures, which can take new employees several months to acquire.

The application uses artificial intelligence to recognise and interpret address information, then guides workers on correct sorting decisions in real time. This reduces dependency on memorised knowledge, accelerates onboarding for new staff, and makes the sorting process both faster and more accurate.
The initiative was developed through Serbia's Innovation Fund GovTech programme, which connects private technology companies with public sector institutions to solve operational challenges using innovative technologies. LambdaWorks, a digital product agency, led product development, covering product management, design, backend and frontend engineering, and infrastructure.

Key outcomes include a greater than 20% increase in manual sorting efficiency and a reduction in employee onboarding time from several months to a few weeks. These improvements directly affect daily processing of 2.1 million plus mail pieces, lowering training costs, accelerating workforce readiness, and yielding tangible financial savings for Serbia's postal infrastructure.

The project also surfaced important lessons. User adoption and change management were significant challenges; introducing an AI tool into frontline workflows required careful attention to usability and trust-building, achieved through intuitive UX design and iterative testing with postal workers. Securing budget for ongoing operation within public sector procurement constraints was a further hurdle. Close collaboration between the technology partner and Pošta Srbije proved essential for aligning the solution with real operational needs, and the GovTech programme model demonstrated its effectiveness as a vehicle for delivering practical AI solutions in the public sector.

Looking ahead, Posta Kopilot has the potential to scale to additional postal facilities across Serbia and could serve as a replicable model for other national postal services facing similar manual sorting challenges. Its modular design supports adaptation to different operational environments.

About the policy initiative


Category:

  • AI policy initiatives, programmes and projects

Initiative type:

  • AI use cases/projects in the public sector

Status:

  • Inactive – initiative complete

Start Year:

  • 2024

End Year:

  • 2024

OECD AI Principles:


Other relevant urls: