TRUSTED DECISION SUPPORT FOR HIGHER EDUCATION

Let Lectora assist in your assessment.

A trusted, grounded workspace for course assessment: your own personal AI teaching assistant, inside Canvas, helping you grade, give feedback and follow up with confidence.

Curious about the results? See how Lectora compares with human graders.

Works with the systems you already run

Lectora installs as an LTI 1.3 tool, inside the course, using your existing roles and roster.

PILOTS AND RESEARCH

See how Lectora has been used in real institutional pilots

See all case studies →

INSIDE LECTORA

Do all your protracted academic work in one place.

Every step powered by AI and backed by statistics, and grounded in your own rubric, course material and learning goals.

01 / GRADE THE SUBMISSIONS

GRADING AND FEEDBACK

Lectora drafts a score and a reason for every submission.

Select the submissions and start the run. Lectora drafts a score and a reason for every criterion on every submission, grounded in your rubric and the course material. Each draft carries a confidence reading, so you know which ones to look at first.

Nothing reaches a student until you save the feedback and publish it to Canvas.

Submissions to grade

LEC207 · Assessment 2 · illustrative

6 selected · ready to grade

Run AI grading
Six submissions in the current assessment run.
SubmissionGradeStatus

Anna Hansen

#A-2104

A
Drafted

Mathias Olsen

#A-2118

B
Drafted

Emma Larsen

#A-2145

D
Drafted

Henrik Andersen

#A-2156

C
Drafted

Nora Johansen

#A-2179

A
Drafted

Ingrid Bakken

#A-2201

F
Drafted

Rubric

4 criteria

Course material

9 files indexed

Drafted

0 / 6

Henrik Andersen

#A-2156

Grade

C

Score

68/100

Confidence64%

Lowest confidence in the run, so it opens first.

Per criterion

  1. Method and setup

    9 / 12

    Regression specified correctly; the factor set is the one the assignment asked for.

  2. Risk adjustment

    5 / 10

    Sharpe ratio computed on monthly returns without annualising. Low confidence: check by hand.

  3. Interpretation

    8 / 10

    Reads the coefficients correctly but does not test their significance.

  4. Presentation

    5 / 8

    Clear tables; two figures lack axis labels.

Written feedback

Right approach, thin evidence. Low confidence: check the risk-adjustment step by hand.

Ready for academic review

Anna Hansen

#A-2104

Draft ready

Grade

A

Score

86/100

Confidence93%

Written feedback

Clear factor comparison. Tighten the interpretation of R².

Ready for academic review

Mathias Olsen

#A-2118

Draft ready

Grade

B

Score

78/100

Confidence89%

Written feedback

Sound method. Recheck annualisation before approval.

Ready for academic review

Emma Larsen

#A-2145

Draft ready

Grade

D

Score

56/100

Confidence85%

Written feedback

Needs review: missing t-tests and weak attribution.

Ready for academic review

Henrik Andersen

#A-2156

Draft ready

Grade

C

Score

68/100

Confidence64%

Written feedback

Right approach, thin evidence. Low confidence: check the risk-adjustment step by hand.

Ready for academic review

Nora Johansen

#A-2179

Draft ready

Grade

A

Score

91/100

Confidence95%

Written feedback

Excellent structure and correct use of the Sharpe ratio throughout.

Ready for academic review

Ingrid Bakken

#A-2201

Draft ready

Grade

F

Score

31/100

Confidence97%

Written feedback

Off-topic: answers a different question than the one set. High confidence this needs a resit conversation.

Ready for academic review

Nothing reaches a student until you save the feedback and publish it to Canvas.

02 / READ THE COHORT

COHORT ANALYTICS

Lectora shows the score distribution, the weak categories, and where each student sits.

Drag the pass line across A–F and watch the split move. Then see performance by category across the cohort, and how one student's profile compares against it.

The distribution reflects graded submissions only. The coverage figure says how many that is.

Cohort analytics

LEC207 · Assessment 2 · illustrative

Drafting progress

114/147

drafts ready

114

in progress

33

Score distribution

N=114

Pass

92

Below threshold

22

Pass threshold

60/100

Pass: 92. Below threshold: 22.

Performance by category

  • Returns82%
  • Sharpe and M²74%
  • Active framing61%
  • Hypothesis tests58%
  • Risk decomposition79%
  • Factor models67%

One student against the cohort

The distribution reflects graded submissions only. The coverage figure says how many that is.

03 / SEE THE COURSE

CURRICULUM AND KNOWLEDGE GRAPH

Lectora shows which of your learning goals the exam does not test.

Each source group holds several files. Select a file for a summary and the related categories. The graph then links those files to learning goals, questions and assessments. Coverage names the goals that still have no approved questions.

Every category and connection the AI proposes is a proposal until an educator confirms it.

Course knowledge

LEC207 · illustrative course

Course sources

Lecture notes

3 files

Used in: Portfolio memo, Final exam

Indexed

Assignment briefs

3 files

Used in: Regression brief

Indexed

Course handbook

3 files

Used in: Portfolio memo, Regression brief, Final exam

Indexed
37 passages

Knowledge graph

  • Categories6
  • Materials9
  • Learning goals5
  • Questions14
  • Assessments3
  • Materials → Categories
  • Materials → Assessments
  • Learning goals → Categories
  • Questions → Categories
  • Questions → Learning goals
  • Questions → Materials
  • Questions → Assessments

Learning-goal coverage

Learning goals
5
Covered
4/5
Gaps
1
Assessments
3

Not yet tested. “Critique model assumptions” has no approved questions in any assessment.

Each number is how many approved questions test that learning goal in that assessment. A dash means none yet.
Learning goalsPortfolio memoRegression briefFinal examBank
Evaluate risk-adjusted returnRisk-adjusted return2125
Test statistical significanceHypothesis testing1—23
Explain factor attributionFactor models—112
Critique model assumptionsModel assumptions———— Gap to review
Compare active and passive strategiesActive framing1124

2Well covered1Thin—Gap

Columns are this course's assessments. Numbers are approved questions touching that goal.

Every category and connection the AI proposes is a proposal until an educator confirms it.

QUIZ STUDIO

Questions written out of the course material you actually teach from.

Quiz Studio drafts multiple choice, short answer and extended response, searching your course knowledge base first under an instruction that every question must be answerable from what the course actually teaches. Each one is tagged to a learning goal your course declared, using only the codes the course itself supplies.

On how good the questions are we claim nothing. There is no study, and the score-agreement result does not transfer to them.

Drafting exam questions →

Quiz Studio

LEC207 · ILLUSTRATIVE

LEC207 / Studio / Canvas

0 questionsNo learning goals selected

Choose a learning goal.

Select a learning goal to open a lane. A category sets scope. Course material focuses the draft.

A question enters an assessment only when a person signs it out, one at a time. Writing a question and signing one out are two different permissions, and an assessment is pinned to the exact question versions it was built from.

Illustrative Quiz Studio question drafting: learning goals, a category and course material added to the canvas, three drafted types, and a chat revision.

INDEPENDENT VALIDATION

Measured on real exams. Published in full, including the study that went against us.

Across twelve sittings of the final clinical exam at the Faculty of Medicine, University of Bergen, Lectora's draft agreed with the course teacher more closely than two independent examiners agreed with each other on the same scripts.

Read the full validation study →

Lectora (0–100) ↑

Course teacher (0–100) →

Lectora vs. course teacher
R² = 0.85
Two human graders
R² = 0.64

12 exams · 889 candidate-exam pairs · 43,683 item-pair comparisons

Dots are illustrative.

CURRICULUM AND KNOWLEDGE BASE

The course material Lectora reads before it drafts anything.

Uploaded material is indexed, split into passages and tagged by category, and learning goals and a taxonomy are derived from it. This is reading, not authoring: Lectora does not write your course, plan your teaching or build your syllabus.

Nothing the AI proposes about your course structure is applied on its own. Categories, goal links and material tags all arrive as proposals with accept and reject beside each one.

What Lectora reads before it drafts →

Proposed category link

LEC207 · ILLUSTRATIVE

Risk decompositionConfidence 87%

This passage defines the decomposition of total risk into systematic and specific components.

“Total variance separates into a systematic term, driven by the factor loading, and a residual specific to the asset.”

Lecture notes, weeks 1–6 · passage 12, p. 34

AcceptReject

When a course has nothing indexed yet, Lectora says so and works from your instructions alone, rather than claiming sources it cannot see.

Illustrative curriculum proposal with a confidence figure, rationale, cited course passage, and accept or reject review states.

COURSE ASSISTANT

Ask about the indexed course material and get the passage it came from.

The assistant answers from the same passages the grading and the knowledge graph read. Every answer numbers its sources, and opening one shows the file, the page and the excerpt it was drawn from.

Read-only in this version. Its Canvas tool surface is filtered to read operations in code, so no prompt or role lets it write to your LMS.

Asking your course material →

Course assistant

LEC207 · ILLUSTRATIVE

Which assumptions does the single-factor model make?

The course treats the factor loading as fixed over the estimation window and the residual as uncorrelated with the market factor.1

1 · Course handbook · passage 4, p. 11

  • Does not grade
  • Does not edit feedback
  • Does not write to Canvas
Illustrative course assistant answer with a numbered course citation and an explicit read-only boundary.

QUESTIONS

The things institutions ask us first.

Read all the questions →

Something else on your mind? Write to us. A person answers.