Teacher Chatbot FAQs

We understand that integrating an AI tool into your course can feel overwhelming and requires thoughtful planning. Below, you’ll find answers to common questions about setting up and managing your Course Chatbot. Each question addresses a specific part to ensure that the adoption process will be as smooth as possible. If your question isn’t answered here, our team is ready to assist you via the Contact form.

Tilburg.ai Chatbot

Our chatbot offers some advantages over commercial alternatives like ChatGPT:

First of all, related to privacy and security: All data remains within the walls of Tilburg University. Instead of sending data to external companies such as OpenAI, we use a copy of a language model (LLM). This keeps all student interactions, and course material within our secure environment.

Second, inclusivity and equal opportunities: We make the chatbot available to every Tilburg University employee and student. This ensures equal access to this assistive technology for everyone within our university community.

Third, remain in control as a teacher of AI usage and responses. You as a teacher will have control over your own chatbot(s). You can easily create or delete a new chatbot via a personal dashboard. You also have the option to add or delete material from the chatbot's "knowledge base". With direct references to your own uploaded documents by the chatbot, students are always updated on the origin of answers. In short: you as a teacher remain fully in control, with students receiving exactly the knowledge that you find important.

The chatbot serves as an extra tool of support in your educational toolbox. By handling routine questions about course logistics and basic concepts, you'll have more time for more detailed discussions and personalized student interaction.

Lecturers notice that thanks to the chatbot, students come to lectures better prepared and ask more in-depth questions. This increases the quality of your contact moments.

As a teacher, you remain in the driver's seat. The information that students receive stems directly from the material that you have entered into the chatbot yourself.

By using the RAG (Retrieval Augmented Generation) technology, students can check the source of each answer. With each response from the chatbot, a PDF can be unfolded that refers directly to the exact location where the information comes from. This allows students to check for themselves whether the information has been provided correctly.

An additional benefit is the insights gained from understanding the questions students frequently ask. You can clearly identify topics causing confusion or gaps in your provided material, enabling targeted improvements in your course content.

Our chatbot offers equal opportunities for all students through a secure educational tool that keeps data within the university, enables flexibility between AI models without vendor lock-in, provides verified answers with direct source references, and saves students from costly subscriptions that would otherwise be necessary for comparable AI support.

Getting Started with Implementation

When you log in with your Tilburg University account, you as a lecturer are immediately recognized as an employee of the university and you get access to a personal dashboard. Within this portal you have the possibility to create your own chatbot.
No, this is not necessary. The chatbot environment is linked to your Tilburg University account. This allows you to log in the same way as you do via Canvas and Osiris.

Managing Content and Access

As soon as a chatbot is created, an invitation link is automatically generated. As a teacher, you can easily copy this link and share it with your students, for example via Canvas or another communication tool. Anyone who has this invitation link and a valid Tilburg University login will be given access to the chatbot. This means that you as a teacher do not have to add separate email addresses to grant students access.
Creating a chatbot is almost instantaneous. You will fill in the necessary details to create a chatbot. After the chatbot has been created, the system still needs some time to process the uploaded documents and make them available. This process usually takes less than 5 minutes. As soon as this processing is complete, the chatbot is immediately ready for use.
At present, it is possible to upload PDF and audio files. In the future, support for Word and PowerPoint files, as well as potentially other formats, will be introduced.
The uploaded files are stored in a Microsoft Azure environment. This keeps the files in a secure environment and they are never shared with the AI model behind the chatbot. Because students have to log in to use the chatbot, only the students involved have access to the uploaded files. If students should not have access to some copyrighted material, please contact us via the contact form.
Yes, it is possible to remove files/course material from existing chatbots.
No, unfortunately it is not yet possible to add new sources later. However, we are working on making this functionality available in the future. We therefore recommend that you think carefully in advance about which sources you want to add to the chatbot.
Yes, you can create as many chatbots as needed. However, please note that usage costs will be charged accordingly. If you have specific applications beyond course chatbots for which you would like to develop a chatbot, please feel free to contact us. We will be happy to explore a tailored solution for your needs.

It is important to be structured when uploading documents so that the chatbot can function properly. Here are some guidelines:


First, think carefully about which sources are relevant. The chatbot can only work with the information that is available. Therefore, make sure that you select all the necessary documents in advance. This can include, for example: weekly lecture slides, the syllabus, articles, papers that are part of the course and practical assignments or practice material.


Second, give files a logical and recognizable name, as the file name plays an important role in how the chatbot processes information. Avoid random names such as "doc1.pdf" or "randomfile.pptx". Instead, use clear names, such as: "Slides_Week_1.pdf", "Syllabus_Course_XYZ.pdf" or "Research_Paper_Smith_2023.pdf". This not only helps with retrieving files, but also with correctly labeling information within the chatbot.


Third, provide sufficient context in the files. The chatbot gets all its information from the uploaded files, so the more detailed and extensive the slides or documents are, the better the chatbot is able to generate relevant answers. Since uploading documents is free, it is advisable to provide sufficient content and context. So don't be too conservative with the material, but at the same time make sure that it remains relevant to the course and ties in with the exam material.