Enhancing Territorial Cohesion with AI: The Castellón Province’s Strategy

This article is based on insights gathered from an interview with Borja Colón, Head of Administration and Public Innovation Service at the Provincial Council of Castellón.

Public administrations are increasingly expected to explore artificial intelligence to improve service delivery, optimise internal processes and respond to citizens’ needs more effectively. However, AI adoption poses specific challenges for small municipalities, where local administrations often lack the technical capacity and organisational resources to deploy it safely and effectively. Better-resourced municipalities progress faster while smaller ones are left behind, creating unequal adoption patterns. This gap does not stop at the institutional level: where citizens and public staff in less advanced municipalities are not equipped to access and trust AI-enabled services, technology may deepen the existing digital divide. Therefore, AI adoption in such territories requires a coordinated strategic response capable of ensuring common standards, territorial cohesion, human-centred governance and equal opportunities to benefit from technological innovation.

With the goal of supporting responsible adoption of AI across the provincial territory, the local governing and administrative body for the province of Castellón in Spain ‘Diputación de Castellón’ has been deploying Provinc.IA, a strategy aimed at driving the territory’s digital transformation through the development, adoption and responsible governance of AI systems that improve the quality of public services, strengthen territorial cohesion, and guarantee the protection of citizens’ fundamental and digital rights, while generating opportunities for sustainable and inclusive socioeconomic development in the area.

The general objective of Provinc.IA is to provide a common framework through which AI can be introduced across the province in a coherent, ethical and scalable way. Rather than forcing  municipalities to experiment independently, the strategy support them through a shared model that can be applied both by the Provincial Council and local entities with different levels of resources or administrative capacity. This is particularly relevant in Castellón, a province composed of 135 municipalities, 90% of which have fewer than 1.000 inhabitants. In this context, the strategy aims to prevent unequal patterns of AI adoption, reduce the risk of digital divides, and ensure that AI-enabled transformation remains aligned with public values, rights protection and territorial cohesion. The foundation of the province’s strategy is the Practical guide on the use of AI in Spanish local entities, developed within the Spanish Federation of Municipalities and Provinces (FEMP) to ensure responsible adoption of AI in local public administrations, combining regulatory guidance, data governance, organisational readiness and concrete municipal use cases.

The strategy builds on the idea that AI should not simply address the shortcomings of earlier phases of digital government. Instead, as declared by Borja Colón, responsible for Provinc.IA at the Castellón Council, 

                                         “It should help redefine the public management model itself and incorporate hyper-automated work dynamics and agentic intelligence,                                                   in order for our organisation to keep up with today’s challenges.”

While previous digitalisation efforts have already tackled issues such as interoperability, organisational silos and technological gaps, AI is seen as an opportunity to go further, enabling higher levels of process automation, administrative simplification and bureaucracy reduction.

To guide this transformation, Provinc.IA is structured around four principles:

1. Human-centred approach, based on human supervision, fundamental rights protection and the principle of keeping people meaningfully involved in AI-enabled processes. 

2. Data governance, recognising that AI is only as effective as the underlying data, as structured and well-managed data flows are the basis for accurate and trustworthy AI-enabled services.

3. Real-world impact, to ensure that AI deployment produces concrete benefits for citizens and improves the functioning of public services. 

4. Stable governance model, designed to make the strategy sustainable over time and capable of adapting to technological and legal changes.

The Provinc.IA model translates its strategic vision into a structured roadmap combining impact and risk assessments, infrastructure deployment, governance and evaluation mechanisms. In the first step the model analyses the state of the province in terms of digital maturity and organisational capacities, by mapping the provincial AI ecosystem, reviewing the applicable regulations and detecting the main challenges and opportunities. Secondly, as stated in the official document of the Artificial Intelligence Strategy for the province of Castellón – Provinc.IA, it establishes 4 strategic objectives:

• Consolidating a provincial ethical and regulatory framework for AI.

• Developing the technological infrastructures and digital capacities needed to deploy AI securely and at scale.

• Promoting organisational change, AI literacy and strategic alliances.

• Reinforcing the social dimension of AI by reducing digital divides and supporting vulnerable groups.

To ensure implementation, the model also includes a monitoring and evaluation system based on strategic and operational indicators, as well as a two-level governance structure: a technical-executive level, led by the Public Administration and Innovation Service and the future Provincial AI Office, and a transversal AI Governance Committee responsible for coordination, impact assessments, audits, shared solutions and the evaluation of new AI systems.

Under the second strategic objective, to translate Provinc.IA from a theoretical policy framework into concrete public-service innovation, the Diputación de Castellón places particular emphasis on the development of technological infrastructure and digital capacities at all administrative levels. This objective is operationalised through the launch of four AI pilots in key public services, which are intended to test how the strategy’s principles can be applied in practice before being consolidated and scaled across the territory. Among them, two projects are being implemented.

The most mature pilot developed within the Provinc.IA strategy is a virtual assistant aimed at supporting citizens in their interactions with the Provincial Council. As described by Borja Colón, the tool combines two channels: a web-based chatbot and a voice interface through a virtual call centre. The objective is to allow citizens to receive support when carrying out procedures, submitting requests or dealing with tax-related matters, including outside traditional office hours, with the goal of completing the totality of the services solely through human-machine interactions. The voice component is particularly significant, as the aim is not only to provide automated answers, but to create a natural and reliable interaction capable of resolving citizens’ requests without immediately redirecting them to a human operator. In this sense, the pilot seeks to improve access to public services while reducing the burden of repetitive front-office tasks.

A second ongoing pilot focuses on a major barrier of transparency portals: the information provided is often written in legalistic and technical language and may not always be accessible to non-specialist audiences. In fact, transparency is insufficient if citizens cannot understand the information made available to them. By making public information more intelligible, the initiative aims to strengthen trust and improve citizen access to administrative knowledge. The tool uses AI to transform legal, administrative and bureaucratic language into clearer language that citizens can understand. This work is supported by methodological guidelines developed with contributions from linguists, data scientists, librarians and legal archivists, setting out steps for converting complex administrative language into clearer formulations.

Looking ahead, while launching the first AI pilots, the priority of 2026 is to develop complementary work on impact assessment, algorithmic transparency and institutional alliances. In parallel, the Digital Literacy and AI Awareness Programme is expected to be operationalised through 22 initiatives over the next three years targeting schools, entrepreneurs and civil society with the aim of enhancing digital literacy across the territory. 

In 2027, the focus is expected to shift from experimentation to consolidation: pilots should become more transversal across the territory, while cybersecurity, interoperability and technological-security risks will become increasingly central. This phase will also coincide with the end of the current electoral cycle, making political continuity an important condition for maintaining the strategy’s momentum. The strategy also foresees the creation of a Provincial AI Office, expected to become fully operational in 2027, to coordinate pilots, standards, monitoring and institutional cooperation. 

By 2028, Provinc.IA aims to have launched a total number of four AI pilots, achieved a 75% pilot success rate, expanded literacy and innovation networks, and strengthened digital inclusion. Beyond that, the strategy’s longer-term ambition is to consolidate Castellón as a living laboratory for responsible AI in public administration, capable of testing, validating and scaling solutions that can benefit both the Provincial Council and the municipalities of the territory.

Provinc.IA offers a range of benefits, especially:

• Strengthen territorial cohesion through a uniform AI model. As planned with Provinc.IA, the objective is to allow AI to be deployed from the smallest to the largest municipality through a uniform process, avoiding different speeds of adoption depending on whether a local administration has more resources or technical capacity.

• Improve services effectiveness and administrative efficiency. AI can help increase processing speed, reduce timeframes, improve the quality of communications and free public employees from repetitive bureaucratic tasks. Citizens should therefore benefit from public services that are faster, more efficient and more human-centred, as staff gain more time to focus on direct support and higher-value interactions.

• Generate economic incentives for the territory. Beyond public administration, the strategy also seeks to support the wider territorial ecosystem by promoting sustainable socio-economic development. It aims to do that by equipping municipalities (especially smaller ones) with shared AI capacities, while also fostering partnerships with universities, GovTech companies and start-ups, as well as research centres and civil society, in order to establish a public-private collaboration ecosystem able to attract talent, technological investment and AI-related R&D projects.

As highlighted by Borja Colón, some challenges have been identified to be addressed in the short and medium term:

• Prevent digital divide by upskilling civil servants and citizens. One of the main challenges is to incorporate AI into public administrations’ internal work and external services without creating negative effects for public servants and citizens, particularly in relation to the required technological skills and unequal access to digital services. To this end, AI literacy initiative will prove fundamental in ensuring administrators and citizens’ readiness.

• Guarantee political continuity beyond the current electoral cycle. Provinc.IA is being implemented in a public-administration context where political support remains essential. As the current electoral cycle ends in summer 2027, the next political leadership will need to validate and sustain the strategy for it to move from pilots to long-term implementation.

• Navigate technological uncertainty. A broader challenge concerns the rapid and unpredictable evolution of AI itself. Colón described public administrations as “navigating on a map that does not exist”, since the future development of AI remains uncertain in both direction and timing. This creates risks around technological choices and long-term planning. However, a model like Provinc.IA grounded in people, ethics, data and stable governance can reduce this uncertainty, because the strategic foundations remain valid even as technologies evolve.

• Manage organisational and cultural change. AI-enabled tools such as a virtual voice assistant or automated document-processing systems can change the daily work of public employees, especially those currently performing repetitive administrative tasks. Thus, staff may either resist these changes. For this reason, the success of Provinc.IA will depend not only on deploying technology, but also on helping the organisation internalise new ways of working to ensure that employees see AI as a tool that supports, rather than simply replaces, their role.

Useful links:

• Estrategia Provinc.IA de Inteligencia Artificial para la Provincia de Castellón: https://www.dipcas.es/uploads/dipcasv2/ficheros/1/6981b7fe962cb2815004396.pdf 
• Guía Práctica y Políticas de uso de la IA en las Entidades Locales (FEMP): http://femp.femp.es/files/566-3598-archivo/IA_EELL_2025.pdf 
• Informe de Diagnóstico Provincial en IA: 68d279c228b954918983005.pdf
• Mapa de Actores y Capacidades en IA: https://www.dipcas.es/uploads/dipcasv2/ficheros/1/68d279dbba1bb8200190523.pdf 
• Identificación de Retos y Oportunidades Estratégicas Para la Provincia: https://www.dipcas.es/uploads/dipcasv2/ficheros/1/68d279eec70b18323332937.pdf 

Project contact:

• Borja COLÓN, Head of Administration and Public Innovation Service, Provincial Council of Castellón – E-mail address: bcolon@dipcas.es