Tomorrow’s technology today!

Our proprietary tool using AI and Machine Learning mechanisms, created on the basis of over 20 years of experience in language level tests and general language proficiency testing (language audits). Based on data collected during nearly 20,000 leveling audits carried out over the past 4 years, we have developed a unique intelligent adaptive language test. It is an innovative tool with the practical application of artificial intelligence mechanisms. The idea of ​​Computerized Adaptive Testing is based on the IRT (Item Response Theory) –  theory of task response using probabilistic mathematical models to express the relationship between the answers to test tasks and the level of knowledge of the tested candidate.

The system matches subsequent questions based on previous answers and the time devoted to them. As a result, the adequate level for a given candidate is reached quickly. Compared to traditional tests, the test time is up to 80% shorter!

How does it work?

The system starts the test with questions on the middle level, and then, depending on the correctness of the answers, it modifies the type of randomly selected questions.
In addition, our system analyzes the time spent on each portion of questions from a given level, adding or subtracting points.
Thanks to this, the system initially diagnoses the approximate level of the tested person and then focuses on confirming it and partial measurement.
Our tests are never the same, there are 47 708 401 986 806 possible combinations of questions in just one test!
We create and constantly modify a huge database of questions for this purpose; currently, IRT Engine draws from over 17 thousand original questions.