4 June, 2019 INTEFbot, a pedagogical virtual assistant with Dialogflow technology INTEF asked themselves how students could make the most of the available resources. The answer was clear: through a technological solution. And there could be no better solution than Dialogflow as a virtual assistant. A space that promotes digital technologies in the educational environment through different plans: training, resources... Yes, that's what INTEF is all about. A website where students can access a wide and heterogeneous range of educational resources. In other words, a website that gives users the possibility of finding the information they are looking for. However, in order for users to easily and accurately find the information they need, it was necessary to identify a system that would really help to locate it easily. Improving access to educational resources INTEF offers educational experiences, digital education, an image bank and a long list of other resources. But the institution realized that something was not working... Users generally had difficulty finding the content they were looking for. And perhaps, when they did find it, it took them several minutes to find it. There was valuable content on the platform that the user probably didn't have access to. So, in the end, the content was available but not accessible. As a solution, Dialogflow met all the requirements: INTEFbot, an educational virtual assistant In this context, the INTEF chatbot was born. A conversational assistant was developed to become a virtual pedagogical agent. Thanks to its technology, it can help users understand the complex INTEF ecosystem. In this way, the main objective was achieved: to help them to easily find the available resources. However, this virtual assistant fulfilled another series of requirements that were fundamental for the institution: Availability in different languages. This chatbot is a project based on Natural Language Processing (NLP) and is integrated with Google's translation API. This makes it possible to generate different conversation scenarios in more than 90 languages. Automatic translation. Originally, this virtual assistant was developed in Spanish. However, the chatbot can automatically detect the language used by the user and translate it if necessary. In addition, the agent's response is also translated so that the user can communicate with the agent. Help detection. The wizard is integrated into the website, and its chat window is activated in different circumstances. One of the times this happens is when the system detects that the user is lost, in other words, when it is looking for a resource. Dynamic responses. Through the use of webhooks and a connection to a database via a firebase application, the wizard can look up specific information and resources for the user directly during the conversation.
