Language-training school ASTEX, based in Madrid, Spain, has seen careers skyrocket after students completed its courses. “Corporate students come to us because their companies need them to improve or update their English skills,” explains Macarena Uceda Márquez, Key Account Manager at ASTEX. “Many of our students need to perform their daily jobs in English, interact with clients, adhere to new American management requirements or prepare to expatriate to an English-speaking country.”

ASTEX students come from various industries and have a wide range of training needs, from seeking intensive immersion to learning sales language around a specific product set. First, students had to complete a written and spoken level test to evaluate their language ability and class level. For the spoken portion, students then booked a phone call with a teacher. On average, the level test was taking students 40–60 minutes.

ASTEX, Ivory and IBM delivered the first MVP in only eight weeks and the second and third MVPs over the next two months. ASTEX developers learned a new programming language and adopted lean, agile ways of working. The team used extreme programming techniques to remove technical debt and change solution features. Agile ceremonies such as weekly iterations, retrospectives and iteration planning meetings helped the team gather feedback from stakeholders and update the solutions as needed.

As a result of the first MVP, the time needed for a student to take a level test has been slashed from an average of 50 minutes to 10 minutes, an 80% improvement. IBM Watson Natural Language Understanding can process students’ spoken and written test answers by extracting and analyzing phrase properties such as number of verbs and nouns.

The second MVP, ASTEX Language Intelligence (Ali), is a digital assistant that accompanies students throughout their ASTEX learning experience. Ali helps onboard students and assists in placing them in the right class because it asks up front what the student’s job is, why they are studying English and how much time they have to devote to language training. The system can interact in a human-like way, helping students express themselves naturally.

Ali also assists with administrative tasks such as canceling a class and finding ways to boost a grade. It provides students with feedback on their progress and can suggest areas of improvement. Ali uses IBM Watson Discovery to recommend exercises that fit with a student’s interests and learning path. For example, if a student has expressed an interest in traveling, Ali will customize that student’s coursework with exercises involving travel so that the content is more engaging. Ali is available to students 24 hours a day, seven days a week.

The third MVP, the digital speaking practice, uses IBM watsonx Assistant, IBM Watson Speech to Text and IBM Watson Text to Speech solutions to help teachers create written or spoken dialogs that vary in complexity based on a student’s needs. This digital tutor and oral language practice are available to students around the clock through a virtual tool or over the telephone.

Because ASTEX team members had such a positive experience with user-centered design, they decided to apply this approach to the ASTEX Language Innovation platform interface as well. Teachers and students gave input on the application’s design and organization, while IBM and ASTEX confidently made changes knowing that they wouldn’t break any other parts of the containerized architecture. The updated design is clearer and more intuitive for users, enabling them to access information and find resources more easily.

User feedback on the improved ASTEX Language Innovation platform has been unanimously positive. Although the core program is the same, the updates seamlessly integrate with the existing system and elevate the entire experience. The platform is more dynamic, with a customized plan for each student. This optimization increases user engagement and improves outcomes.

The learning platform’s AI-driven components are revolutionary. “No one else in the language-training industry offers digital speaking practice like this,” explains Márquez. “The assistant Ali is also a differentiator because people have become used to chatbots, but it is a shocking difference to actually be able to have this kind of interaction with a machine.” Because of her machine-learning capabilities, Ali will continue to become even more intelligent with every input. Further, since technology automates simple tasks, teachers can focus on more complicated tasks—and teaching.

ASTEX is now able to easily scale its platform, which Ivory and IBM plan to help evolve. “At ASTEX we have always been clear that being at the forefront of technology and pedagogical innovation is key to maintaining ourselves as a benchmark in the sector, and that is why we rely on the best and invest heavily in constant evolution,” explains Alfonso de la Torre, Director of ASTEX. “In this case, who better than IBM to accompany us to apply Artificial Intelligence to language training?”

Ready to see what IBM Garage can do for your business? Talk to an IBM Garage expert.