A shared standard.
Keep the rubric visible across every answer. Make differences in interpretation easier to identify and discuss.
AI for assessment
AI-powered assessment intelligence
for modern educational institutions.
The process
Evaluate, explain, and verify student answers with clarity and consistency.

Different
subjects
same
clarity
01 — UnderstandBring questions, answers, and their academic context into one clear view.
Built for real education
From the first response to the bigger picture.
A connected approach to academic assessment.
Designed for education
From university exams to professional certifications, assessment starts with your subject and criteria.
INSIDE THE INTELLIGENCE
[ A CONNECTED APPROACH ]Assessment is a chain of decisions. KRITIS is being developed to make that chain more visible, from the first question to the final review.
Connect passages in an answer to the relevant assessment criteria, so the basis for a decision stays visible.
Drag to rotate · Select a node to explore
DESIGNED AROUND EDUCATORS
Intelligence should support the people responsible for learning. These principles guide the KRITIS assessment workflow.
Keep the rubric visible across every answer. Make differences in interpretation easier to identify and discuss.
Connect decisions to answer evidence, with explanations that invite scrutiny and correction.
Shape assessment around the subject, the question, and your institution’s marking approach.
A workflow designed to reduce repetitive review, so educators can focus on the decisions that need their expertise.
Educators review findings, resolve exceptions, and retain responsibility for final assessment decisions.
Agree on access, retention, and institutional requirements before an evaluation uses real student information.
RESEARCH & VALIDATION
Useful assessment intelligence needs more than a compelling demonstration. It needs clear methods, real academic scrutiny, and honest limits.
VALIDATION FRAMEWORK / PROPOSEDCompare assessments against explicit rubrics and independently reviewed reference answers. Examine disagreements by criterion, subject, and answer type, including where the system should defer to an educator.
Proposed methodology · Results are not yet published here.Subject experts should define the rubric, inspect evidence, review edge cases, and retain the final decision. The goal is a reviewable workflow that supports professional judgement.
Design principle · No expert endorsements are claimed.No independently verified performance metrics, published studies, or institutional partnerships are presented on this website. Published evaluation evidence will be needed to establish model performance.
Research status · Evidence will be published when available.A proposed pilot would begin with agreed evaluation goals, authorized sample material, human review, and institution-specific data handling requirements. Scope and success criteria should be settled before student data is used.
Record your interestTHE KRITIS PERSPECTIVE
It is a window into how someone thinks. Assessment should make that understanding clearer, for educators and learners alike.
KRITIS explores how computational intelligence can support that work: connecting answers, criteria, and explanations while keeping academic judgement with people.
START WITH A CONVERSATION
Tell us about your institution and the assessment challenges you want to explore.
Secure request collectionRequests are stored on the KRITIS server for the team to review. No automatic confirmation email is sent.