Breaking down silos in the Italian public sector

Formez PA is the technical agency of the Department of Public Administration in Italy, and it specializes in public sector transformation and modernization processes. The agency identified a challenge that was prevalent not just in the public sector in Italy but was also true across other nonpublic institutions globally. Often, public institutions and the departments within them operated in silos, each with its distinct professional profiles and terminologies. This lack of standardization, coupled with an outdated approach to human resource management that is not conducive to organizational flexibility and professional development, stood in the way of effective collaboration and data exchange, causing confusion.

Formez PA identified that in the absence of a consolidated framework that was accepted across institutions, the widespread use of different terminologies to categorize professional profiles complicated the comparison and integration of these profiles. Consequently, inter-agency collaboration became difficult, often resulting in service delivery delays and slowing down teams.

This lack of standardization also resulted in data inconsistency. Discrepancies in terminologies across institutions led to inaccuracies in data collection, analysis and interpretation, thereby compromising the quality of decision-making processes. These inconsistencies highlighted the need for a comprehensive solution that could uniformly classify professional profiles and include a set of skills—also belonging to a single dictionary—for each profile. Such a solution would promote smoother connections between diverse public administrations. Formez PA thus looked to build a solution that could unify and standardize processes, terminologies profiles and skills—to enhance inter-agency collaboration and service delivery efficiency across Italy’s government entities.

Formez PA was entrusted with developing a practical AI solution to classify professional profiles in the public sector effectively. Formez PA collaborated with IBM® Client Engineering to design an advanced AI-driven solution integrated into the existing RiVa portal. The general objective of RiVa is the design, construction, experimentation and development of a competency-based Human Resource Management (HRM) model for the implementation of strategic planning and management processes, valid for all public administrations and adapted to the specificities of different sectors. The ongoing collaborative project aims to identify a unified and integrated methodology for strategic human resource management based on competency models as well as promote the evolution of the labor market and career development within public administration. AI can help the Department of Public Administration classify professional profiles and standardize professional terminologies, fostering a unified understanding across the department.

The solution leverages the IBM watsonx® portfolio of products to translate user queries into embeddings. Using the Llama large language model (LLM) in IBM® watsonx.ai™, the solution enables the comprehension and interpretation of natural language requests. This is complemented by IBM® watsonx.data™, which leverages Milvus as its vector database to execute semantic searches within vector spaces. This functionality enables the solution to pinpoint professional profiles that share semantic similarities, even when the keywords don't exactly match.

To facilitate seamless integration with the RiVa portal, the IBM–Formez team devised a microservice to expose APIs. This service doubles as an intermediary, invoking watsonx.ai to convert user requests into embeddings and using watsonx.data for conducting semantic searches.

The team further enhanced user-centricity by implementing filters based on contractual areas, professional dimensions and professional families, empowering users to home in on specific search criteria. Today, the system can suggest profile names, organizational purpose and primary activities based on the user's description. In future iterations, this feature is expected to evolve to recommend skills and role profiles, further enhancing the user experience. In the spirit of continuous improvement and user-centric development, the solution also has a built-in feedback mechanism for gathering user inputs to refine the solution’s performance iteratively over time.

The solution presents a robust and user-friendly approach to professional profiling, and it is expected to be a unique touchpoint for professional profiles in the public sector.  By leveraging IBM's AI capabilities and Formez PA's expertise in public sector projects, the solution targets the standardization of professional terminologies, professional profiles and skills across the public administration landscape.

One of the key benefits of integrating AI into RiVa is the potential of the technology to significantly enhance productivity. The solution employs IBM watsonx.ai to generate professional profiles via natural language input, and Formez PA anticipates approximately 90% reduction in manual data entry and data processing. This reduction will not only save time but also free up resources for more strategic initiatives within public administration.

Another critical advantage of the solution is its improved search accuracy. Through the integration of semantic search functionalities offered by IBM watsonx.ai and watsonx.data, public bodies can expect approximately 80% increase in search precision. This enhancement transcends the limitations of traditional keyword-based searches, enabling the solution to comprehend semantic nuances and highlight the most pertinent professional profiles.

The user experience is also set to improve markedly, with an anticipated increase of up to 60% in user satisfaction. This improvement stems from the solution's faster and more accurate search capabilities, enabling public administrations to connect professionals with suitable roles more efficiently.

By standardizing terminologies, minimizing manual input, enhancing search accuracy and elevating user satisfaction, the infusion of AI into the RiVa portal is expected to redefine operational efficiency in human resource management in the public sector.

Formez PA founded in 1998, is the technical agency of the Department of Public Administration in Italy, and it is based in Rome, Naples and Cagliari. Serving public administrations, local authorities and public enterprises, it offers strategic solutions in organizational development, training and service innovation. The agency mission is to support public administrations in achieving their strategic objectives, enhancing service quality and promoting efficiency in the delivery of public services.