AI teaching assistants Inside Moodle: Why universities are rethinking student support

Students are already using AI while studying, and we are increasingly discussing with universities that are wondering where these interactions should happen. In many courses today, we notice the same pattern:

• a student is working on an assignment late in the evening and runs into a problem;

• instead of waiting for office hours or posting on a discussion forum, they open ChatGPT in another tab, paste the question, and get an answer instantly;

• from the student’s perspective, this is convenient. But from the university’s perspective, it creates a gap.

Obviously, the AI has no awareness about course context, nor the professor’s framework or the materials that were carefully prepared inside the LMS. Reciprocally, faculties can’t currently access these “shadowGPT” interactions, and institutions lose visibility into how students are actually learning. Over time, there is a real risk learning experience starts drifting away from the platform where the course itself lives.

If we go down this line, there is a tangible risk academic integrity cannot be granted; and that university degrees value’ ultimately decrease because of AI.

The external AI tool problem

Many universities have started experimenting with AI tools in recent years. Some focus on proctoring, others on analytics or automated grading. But when AI exists outside the LMS, it often adds another layer of complexity rather than solving the real problem. Indeed:

1. students have to move between multiple tools;

2. faculty loses insight into where students struggle;

3. dearning data ends up scattered across different platforms.

For institutions dealing with data governance requirements like GDPR or FERPA, external AI tools can also introduce questions about where conversations are stored and how that information is handled. The challenge isn’t that AI exists and students will continue to use it. It is rather that most AI interactions currently happen outside the learning environment universities control.

Bringing AI into the LMS

A different approach is starting to emerge: embedding AI directly inside the LMS itself.

When an AI assistant is plugged inside of a given LMS (like Moodle, Blackboard, Canvas), it can understand the underlying course structure. It knows which module a student is studying, which materials belong to that lesson, and where the student stands in her learning journey. Instead of asking a generic AI tool for help, students can ask questions directly within their course environment. The answers reference the course content and guide them back to the materials their professor designed. This keeps learning support connected to the course rather than disconnected from it.

For faculties, embedding also brings useful visibility. Over time, professors and learning teams can see which concepts generate the most questions, which resources students struggle with, and where additional explanation might help. In large lecture courses, where a single professor might be supporting hundreds of students, this kind of assistance can make a meaningful difference.

Making learning adaptive, without replacing teaching

The endgame for educational AI in higher education isn’t to replace professors; and at Raison, we believe teaching is fundamentally human.

Rather, the real opportunity lies in moments where traditional support simply doesn’t scale; especially when students are studying late at night, preparing assignments, or reviewing complex concepts. An embedded AI assistant can provide immediate clarification during those moments while still keeping the learning process connected to the course.

Where Raison fits in

At Raison, we are building an AI teaching assistant designed specifically for any LMS.

Rather than creating another external platform, our goal is to embed AI directly inside the LMS so that students receive contextual help within their courses while institutions maintain visibility and control over the learning environment.

AI will inevitably become part of higher education. The important question is not whether students will use AI or they already are. The real decision universities face is whether that interaction happens inside their learning ecosystem, or somewhere outside of it.