At Amsterdam Tech, our online bachelor programmes in Software Engineering, Data Science and Machine Learning are built around hands-on, project-based learning. Students can study from anywhere and organise their learning around work and other responsibilities.

This flexibility also means that students sometimes need support when no human mentor is available. They may get stuck while coding, need clarification after missing a live session or want to check their understanding before continuing. We built the AI Mentor to support them in those moments as part of Amsterdam Tech’s innovative AI-supported learning model.

Why We Built It

General-purpose AI tools can answer technical questions, but they rarely understand the student’s programme, project requirements or learning objectives. They may generate a solution without recognizing what the student is expected to learn, and may complete the task instead of helping the student think through it.

AI Mentor was designed to provide immediate, context-aware support inside Amsterdam Tech’s learning environment. It helps students clarify requirements, work through problems and continue independently while protecting learner autonomy, strategic thinking and academic integrity.

What the Experience Is Like

AI Mentor is a real-time, multimodal mentoring environment embedded in the ecosystem. Students speak or write to a responsive video persona and continue the conversation as they work.

The experience is conversational and dialogical. Students explain what they are trying to do, describe what they have already attempted and respond to follow-up questions. The mentor adapts its guidance as the conversation develops.

This makes the interaction closer to a human-like coaching conversation than a conventional text-based chatbot exchange. Students can talk through a problem, examine their assumptions and make their reasoning visible instead of submitting one prompt and receiving a finished answer.

How It Works

AI Mentor is grounded in curated Amsterdam Tech curriculum and project materials. It responds in relation to the student’s actual project, learning objectives and programme expectations.

Its approach is based on Socratic coaching. Rather than immediately providing a solution, the mentor asks students to explain requirements in their own words, identify where a problem begins, compare possible approaches, predict the effect of a change and decide what to test next.

Pedagogical guardrails prevent it from completing assessed work or providing complete final code. Its role is to coach students through the problem while keeping them responsible for their own decisions and solutions.

The interface supports real-time video and voice interaction, text responses, live transcripts, visible on-screen questions, downloadable conversations and session reconnection. Interactions may also create reviewable, low-stakes learning evidence after a missed live session.

The experience has been developed iteratively with students, whose feedback has shaped both the interface and the mentor’s educational behaviour.

What’s Next

The beta has already generated sustained, multi-turn learning conversations. Students rated the mentor an average of 4.5 out of 5 for helping them understand their topic.

We will continue expanding AI Mentor across more projects and programmes while evaluating reliability, learning impact and responsible use. Human faculty will remain responsible for deeper coaching, assessment and academic judgement.