AI tutor for universities, trained on your own courses

Because a generic chatbot answers with what it learned from the internet: it can contradict the professor, invent references or solve the whole assignment. The Geedle tutor uses RAG over the course materials, which means it first searches the readings, slides and recordings the instructor uploaded and only then writes the answer, citing the source. If the question falls outside the syllabus, it says so and suggests asking the professor. That way AI in higher education reinforces what is taught in class instead of competing with it, and every answer can be checked.

Each course has an academic integrity mode the instructor configures. In that mode the tutor does not hand over answers to homework, exams or projects: it explains the concept, asks guiding questions, proposes a similar exercise and points to the part of the lecture where the topic was covered. The instructor decides which activities are locked and on which dates, for example during an online exam. Conversations are logged for coordinators according to university policy, and students know from the start that they are talking to an AI and what its rules are.

Instructors stop answering the same question by email twenty times and receive a summary of the topics that confuse their group most before the exam. Academic coordinators see, by course and program, which content generates the most questions, when students study and which materials should be redone. To prepare faculty for this new model we offer AI teacher training, and if the university does not yet have a solid virtual campus, we implement one with our Moodle-based e-learning platform, with the tutor integrated from day one and no complicated migrations.