ThesisMentor AI is a specialized AI Agent platform for the mentorship and assessment of graduation theses at universities. Unlike a one-off grading assistant, it models the entire thesis journey as **seven standardised phases** — from topic selection and literature review to methodology, implementation, defence preparation, and review.
One platform that supports **all four roles** in the academic process: students, supervising advisors, independent examiners, and academic administrators. AI serves people, preserves academic value, and always keeps the final decision in the hands of the expert.
ThesisMentor AI is not a general-purpose chatbot dressed up for education. Each capability below was designed for a specific situation in the thesis supervision and assessment process at a university.
Each phase — from proposal, literature review, methodology, implementation, to defence — has clear completion criteria, expected artefacts, and an explicit advisor approval step. Students always know where they are and what to do next.
Every advisor has their own assessment criteria, citation standards, and feedback style. The AI learns those settings instead of imposing one template on every student.
First-class support for the examiner role: an independent grading workflow that does not see the supervising advisor's comments — preserving the objectivity of the review round.
AI plays the part of committee chair, methodologist, and domain expert across multiple rounds of questioning — letting students rehearse their defence well before the real committee.
The system tracks several risk patterns — schedule slippage, weak argumentation, missing citations, scope drift, methodology mismatch — early enough that the advisor can still meaningfully intervene.
Every AI interaction by a student is recorded and auditable, and every AI suggestion links back to the specific passage in the draft — supporting academic integrity policy while still allowing responsible AI use.
A personal workspace per thesis: the seven-phase roadmap, weekly updates, an AI co-pilot during writing, and a defence simulation before facing the real committee.
A central inbox for every phase approval request, a personalised assessment rubric reflecting the advisor's own style, and early-warning signals across every student under supervision.
An independent examination workflow — isolated from the advisor's comments — with structured reports that preserve the objectivity of the review round.
Aggregate dashboards covering supervision progress and quality across departments, programs, and the whole institution, with support for running multiple departments in parallel.
Six core surfaces of the platform: the seven-phase thesis journey, the advisor's per-phase approval inbox, the six-rule risk signal report, the mock defence with three committee personas, an AI Coach chat that helps build the proposal, and an independent examiner scoring view. The carousel auto-advances; hover to pause.
7-phase roadmap · Semester 2 — 2025–2026
P1
Topic
P2
Literature
P3
Methodology
P4
Proposal
P5
Implementation
P6
Discussion
P7
Defence
Current phase
Define research methods, baselines, evaluation plan and submit for advisor review.
We are looking for university partners to deploy ThesisMentor AI in the coming semester.
