Autoři

Lesia Hrechukha, Nataliia Butko

Anotace

EdQuality AI addresses the difficulty of navigating dense university documentation. Theproposed no-code assistant uses a Google Gem grounded in official educational programmes,syllabi and normative documents to answer questions for applicants, students and academicstaff. It is intended to clarify programme structures, competencies, electives, career optionsand administrative information while directing users back to source sections. The project alsoexplores personal assistance for departmental document work and accreditation-relatedinformation retrieval, with human verification retained for consequential uses.

The presentation framed the problem as one of access rather than document availability.University websites offer static PDF downloads and keyword search, while users mayotherwise consult staff or ask general chatbots such as ChatGPT or Gemini. The team arguedthat official material is difficult to read because of specialized academic vocabulary anddispersed information: applicants struggle to identify career options, students may havedifficulty selecting electives, and staff spend time cross-checking syllabi. General-purposesystems were described as insufficiently localized and capable of producing incorrect answers.

EdQuality AI was presented as an intelligent layer over authoritative documentation. Officialeducational programmes and syllabi are supplied as source data; Google Gems retrievesrelevant content; strict system instructions guide generation; and the assistant returns acontextual response. The presenter said that output can cite relevant sections so users caninspect the basis of an answer. The prototype is specialized for comprehensive normativedocumentation associated with a bachelor’s-level applied linguistics education programme. Its interface and content were shown in Ukrainian because it was prepared for use within theteam’s department.

A live screen demonstration showed the working assistant accepting a request and returning“3 years and 10 months” as an answer. The presenter acknowledged that processing can takenoticeable time, although it still reduced manual searching in their experience. Access iscurrently better suited to personal or limited-group use: adding a user requires knowing anemail address, and Gemini/Gem sharing imposes practical limits. For sensitive tasks, includingaccreditation-related document retrieval, the assistant is positioned as a search aid rather thanan autonomous auditor; its outputs must be rechecked.

licence

Creative Commons BY-SA 4.0