One of the Policy Innovation Lab’s publicly available tools is its catalogue of AI-powered tools. The catalogue illustrates how governments worldwide utilise AI. This December, we have again updated our catalogue to ensure it remains relevant and up to date, as governments incorporate rapidly changing AI technology into their service provision.
As before, this update is based on the latest release of the Public Sector Tech Watch database. This is a database of tools that utilise emerging technologies, found in the European Union’s public sector. We filter this database to ensure relevance to South Africa and narrow our focus to AI-based applications. We then extend the data based on our own findings and research to include AI-powered tools from other countries. We found that South Africa has a large number of applications in security and law enforcement relative to the EU. In total, this end-of-year update includes over one hundred new examples of AI in government from around the world.
An interesting trend has been the continued relative increase in the number of AI-powered procurement tools. This follows many previous examples of AI used to detect fraud based on financial transactions (for example, BRAVA in Spain). However, Large Language Models and agentic systems are now used to automate aspects of the procurement process and to explain procurement legislation and processes to businesses. For example, in Albania, a procurement platform will automate procurement processes using AI to draft terms of reference, estimate costs, and generate initial evaluation reports, thereby increasing transparency and reducing human errors while maintaining human-in-the-loop for oversight.
Another trend is the increase in cross-government AI-enabled data platforms. For example, PLAIN (Platform Analysis and Information System) is a cross-departmental data infrastructure that lays the technological foundation for future AI and data applications in the German Federal Administration, while using AI to perform routine data quality checks. Similarly, Luxembourg uses AI optical character recognition models to digitise historical archives. Ukraine’s Diia platform also illustrates this trend: although primarily a digital identity and public service ecosystem, it integrates AI for automated document processing, fraud and anomaly detection, and AI-assisted citizen support, demonstrating how large-scale digital platforms can embed AI directly into service delivery while enabling wider AI adoption across government.
A key question raised by these examples is how to strategically adopt AI so that it unlocks further adoption of AI. AI-enabled data infrastructures provide greater access to high-quality, timely and large datasets, allowing further AI development. Incorporating AI into procurement processes may enable more AI adoption, since difficult procurement systems often function as barriers to incorporating AI-powered digital technologies into government in the first place. This is particularly true in South Africa, where digital transformation is frequently hindered by procurement processes, which the government is actively trying to improve. At the Policy Innovation Lab, we are in the process of strategically incorporating responsible AI into South Africa’s policymaking process.