- Open administration
New recommendation guides to improve the implementation of AOC services
The Open Administration of Catalonia (AOC) considers the user support service as a fundamental piece to progress in the deployment of digital services and digital transformation. It currently invests more than 1.5 million euros, but it is not enough to meet all the demands of the public and the staff of the Administration.
With the available resources, the AOC has sought innovative solutions to extend the scope of support. In 2019, it launched its first chatbot using artificial intelligence algorithms. This virtual assistant had the advantage of being operational 24 hours a day, with a low cost, but it was only able to answer simple queries and, to achieve a good quality of answers, required a significant training and maintenance effort .
Now the AOC has taken a qualitative leap in its user support service with the implementation of a state-of-the-art chatbot based on generative artificial intelligence (AI). In the field of attention to citizens, this is a leading solution in the State and comparable to the most advanced initiatives on an international scale.
Now available in four services
The chatbot is currently operational in four key AOC services: idCAT Certificat, idCAT Mobile, e-Notum and Valid. This system offers several significant advantages over traditional models: greater understanding of natural language; more relevant and personalized responses; and continuous adaptation to new queries, reducing training and maintenance effort.
The chatbot uses a generative artificial intelligence engine that processes natural language and generates clear, concise and relevant responses. The chatbot's knowledge base comes from frequently asked questions (FAQs) and recurring support tickets, and is continuously adapted to new queries through a feedback system.
With high scalability the chatbot is ready to adapt to new demands and expand to more services, offering multi-channel integration and multilingual support. However, it still faces challenges such as accuracy and maintenance of information, as well as handling complex or critical queries.
88% hit rate
Since its launch in May 2024, the chatbot has assisted an average of 11.575 people per month, becoming a recurring and effective consultation tool. With a hit rate of 88%, the system has proven to be able to satisfactorily resolve a wide variety of queries, reducing the number of queries received by the AOC's User Support team and reducing waiting times .
We're still learning its potential, but the AOC's generative AI support chatbot is already proving to be an excellent example of how technology can improve the efficiency and effectiveness of public services, providing a quality support to citizens and optimizing available resources. With its ability to scale and adapt, this project has the potential to continue to grow and improve in the future, offering more and more services and more accurate answers.
In addition, the project has proven to be a key tool for the continuous improvement of AOC services. Its ability to adapt to the changing needs of citizens and its effectiveness in resolving queries have been determining factors in its success.
90% savings
In a preliminary analysis we estimated that the use of generative AI chatbots generates a 90% saving in attending to a user query. The average cost of human attention is approximately €15, while questions resolved by the chatbot cost around €1,5.
This saving will allow the AOC to dedicate more resources to accompanying people with a lack of digital skills and more complex requests in a personalized way.
Ensuring ethical and responsible use
The AOC is committed to ensuring ethical and responsible use of generative AI. This involves establishing clear policies on privacy and data security, ensuring that the responses generated are appropriate at all levels, while maintaining a firm commitment to user rights.
A thorough risk analysis has also been done to identify and mitigate potential issues associated with the use of generative AI. This includes the assessment of technical, operational and ethical risks, as well as the implementation of preventive measures to minimize these risks.
In addition, the Support team performs daily human supervision of the conversations, and this allows for a constant update of the knowledge base and ensuring that the chatbot is always ready to offer the best possible answers.
The reflection of all this culminates in the Algorithmic transparency sheet, to which the AOC shares all the technical, legal and motivational details of the project, and which we encourage you to consult.
This initiative is the result of the joint work of the Support and Innovation teams of the AOC, and the collaboration of the company 1MillionBot.