Assessments moved onto the portal — pupils complete all five methodologies online, at their own pace, with automatic scoring.
The interview stayed human — a specialist still holds the conversation in person, but the write-up now enters the system as a first-class input alongside the test results.
Synthesis went to the model, under Talim Pro's rules — the OpenAI API combines assessment results and interview text into a finished report, following the interpretation logic and structure defined by Talim Pro's own methodologists.
Seven roles matching the real participants — student, parents, form teacher, principal, district education head, Talim Pro specialist, and Talim Pro admin, each with their own view.
Bilingual by design — Russian and Kazakh across the interface, the assessments, and the generated report itself.