Where tutor judgment fits
See how Axtant checks AI marking, routes exceptions to tutors and keeps the source work beside each decision.
The guide sets the marking basis
Axtant connects each written question part with its mark value and relevant guide. Missing or conflicting marking authority needs attention. A question-derived basis, where supported, is identified explicitly rather than being presented as your supplied rubric.
Checks before a result proceeds
Axtant marks each written answer several times independently and checks that the result belongs to the right question part and stays within its marks. An answer waits for a tutor when the markings disagree, the writing is hard to read, a calculation needs checking or Axtant cannot tell where the answer is. After release, tutors can spot-check a few automatically released marks chosen at random.
These checks reduce avoidable errors. They cannot establish that every AI judgment is correct. Your agency remains responsible for how marks are used.
Review the work beside the decision
Tutors see the submitted page beside the suggested mark and its reason, then use the mark or adjust it. Corrections stay with the marking record. The review queue holds only the answers waiting for a decision.
Keep room for your teaching judgment
Essay evaluation, unusual working and incomplete responses can require a teacher’s interpretation. Axtant’s examples illustrate specific guides and responses; they are not an accuracy benchmark or a guarantee about your worksheets.
Use the cost calculator to estimate monthly savings while allowing for tutor checks and exceptions.