Recently, the AOC attended the Gartner Conference, organized by the Telecommunications and Information Technologies Center of the Generalitat de Catalunya, where the executive conclusions of the prestigious Gartner® IT Symposium/Xpo held in Barcelona in November 2025.
At a time when Artificial Intelligence (AI) seems to be flooding everything, the day has served to put our feet on the ground: only 5% of organizations are getting real benefit from AI, despite the fact that 20% already use it.
What differentiates this 5% from the rest? What challenges await us between now and 2030? At the AOC, we analyze the main lessons for the public sector.
1. Strategy before algorithm
95% of AI pilot projects are unsuccessful. The reason is not technological, but management: lack of a roadmap, vague objectives and absence of evaluation metrics. For public administration, value should not be measured only in internal productivity, but in three axes:
- Social return: real improvement in the life of the citizen and the experience of the public employee.
- Economic return: efficient cost management (including “hidden” costs).
- Future return: preparing the bases for technological sovereignty.
2. Mature technology, people in transition
One in four AI errors is due to poor data quality or poor training. But the biggest risk is theatrophy of knowledgeThe technology is ready, but the organizations are not yet.
Robust and automated governance is needed to keep errors to a minimum (ideally, less than 0,01%). AI adoption only works when teams are mixed (human and machine) and when the employee does not feel threatened. The data is clear: only 1% of actual layoffs are directly due to AI. Success lies in demonstrating that AI does not replace people, but rather reorients roles towards higher-value tasks.
3. The era of agents and digital sovereignty
Looking ahead to 2030, the AI agents will be the main protagonists, deploying quickly to take on complex operational tasks, although there is little real experience at the moment. However, this carries a risk: the vendor lock-in (dependence on suppliers).
In light of this, the European Union is already leading the way:
- Digital Sovereignty Plan: All states will have to have one by the end of 2027.
- OpenCloud Initiative: An effort to cover everything from hardware to software and ensure that we do not lose control of our critical infrastructures.
4. A new regulatory framework: The EU Omnibus Law
One of the big news is the upcoming approval of the Omnibus Law on Regulatory SimplificationThis law, which is expected to be approved in the coming months, aims to relax some GDPR obligations to eliminate overregulation and gain competitiveness.
This will open the door to technologies that were previously very restricted, such as passive biometrics, facilitating a more agile relationship between citizens and administration.
5. Towards a culture of tenacity
Disruptive innovation requires a profound cultural shift. Gartner emphasizes that organizations must reward profiles with three key characteristics:
- Agility: to adapt to constant changes.
- Risk assumption: tolerating error as part of learning.
- Tenacity: the ability to get up quickly, review what went wrong and continue.
Looking to 2030: New horizons for public service
The future holds radical changes that will transform our relationship with technology. In five years, AI will allow us to achieve challenges that were previously unthinkable: preventive detection of fraud on a massive scale, automation of citizen response, proactive personalization of social services or automation of repetitive tasks (elaborating standard reports, translations, signature management, classification of applications or files, etc.).
In the near future we will stop doing jobs that will no longer make sense in an automated environment.
Conclusion
The message for the administration is clear: the real value of AI is not obtained by reducing staff, but by changing the mindset. Success depends on an ambitious strategy, impeccable data governance and the ability to reorient human roles to accelerate processes and generate new public value for the citizen.
