03 / How it works
One assignment, followed from audit to assurance.
CoAcademic is a review instrument, not an autopilot. The journey is designed so that a faculty member can follow any outcome claim backwards through the Learning Evidence Chain — and break it wherever they disagree.
01The journey
Audit → Protocol → Calibration → Evidence review → Assure
Most of this journey is founding-pilot work carried out alongside a course team. Only the evidence review surface exists today as an illustrative prototype.
01Audit
Audit the assignment for exposure
Start with one real assignment. The Assessment Resilience Audit reads the task against its intended outcome and asks where a generative model could satisfy the requirement without the learning happening — and which part of the task actually carries the outcome.
Assessment Resilience Audit · pilot scope- Intended outcome
- confirmed
- Outcome-carrying task
- identified
- Exposure points
- listed
- Detection used
- never
02Protocol
Set the AI-use rule and the evidence plan
Decide whether AI use is restricted, limited, disclosed, or required for this task — any of the four can be valid when it is aligned to the outcome. The Evidence Plan then states which process artifacts, contextual application, and explanation the assignment will require, before students begin.
AI Use Protocol · Evidence Plan- AI-use mode
- declared
- Process evidence
- specified
- Explanation prompt
- written
- Disclosure format
- agreed
03Calibration
Agree what each band means
Prior graded examples and anchor work fix band interpretation with the faculty who will mark. Where several markers are involved, calibration is what makes their judgments comparable — and disagreement is surfaced rather than averaged away.
Rubric Calibration · pilot scope- Rubric version
- confirmed
- Anchor work
- faculty supplied
- Band meaning
- agreed
- Marker variance
- surfaced
04Evidence review
Read the evidence behind the reading
Each criterion arrives with the excerpt it rests on, the calibrated band language, a written rationale, and a statement of uncertainty — alongside the student's own explanation. Faculty approve, revise, or reject per criterion, and results stay in draft until a named educator releases them.
Evidence-Linked Review · prototype- Excerpt p. 4 · ¶2
- linked
- Student explanation
- attached
- Proposed band
- 4 / 5
- Released to students
- not yet
05Assure
Preserve the record
What remains is the Assurance Record: outcome, requirement, AI-use rule, required evidence, artifact, explanation, rubric version and calibration, and every faculty decision — held together so the outcome claim can be inspected later rather than simply trusted.
Assurance Record · pilot scope- Chain links
- 01–09
- Rubric version
- recorded
- Faculty changes
- recorded
- Accreditation reporting
- not offered
02The Learning Evidence Chain
Nine links behind one outcome claim
Product, process, and explanation together. Each link contributes evidence; none of them proves authorship on its own.
- 01
Intended learning outcome
The outcome this assignment is meant to evidence, stated before anything is designed.
- 02
Assignment requirement
What the task asks for, and which part of it carries the outcome.
- 03
Permitted AI use
Restricted, limited, disclosed, or required — declared in advance and aligned to the outcome.
- 04
Required process evidence
The planning, drafting, or annotation artifacts the task asks students to keep and submit.
- 05
Student artifact
The submitted work itself — one link in the chain, never the whole of it.
- 06
Explanation or defense
The student's account of specific choices, in writing or a short targeted defense.
- 07
Rubric judgment
A criterion reading measured against calibrated band language, with the excerpt attached.
- 08
Faculty decision
Approve, revise, or reject. The educator decides, and nothing releases itself.
- 09
Assurance record
The chain preserved end to end, so an outcome claim can be inspected rather than trusted.
Each link contributes evidence. No single link — not the artifact, not a process log, not a local prompt — proves authorship on its own.
03Moderation
A second reader, where teams mark together
Founding pilot scope
Where a department marks in teams, consistency is the real problem — not speed. The moderation pass is founding-pilot scope, not a shipped module.
- a
Sample selection
A subset of scripts — boundaries, outliers, and a random sample — is routed for a second reading.
- b
Independent second judgment
The second marker sees the rubric and the script, not the first marker's decision, until they have recorded their own.
- c
Disagreement made visible
Where markers differ, the difference is shown and discussed rather than averaged into a number nobody chose.
- d
Recorded resolution
The agreed outcome and the reasoning behind it become part of the assessment record.
04Boundaries
What the workflow will not do
No final grade
CoAcademic proposes bands. It does not issue, calculate, or commit a final grade.
No automatic release
Nothing reaches a student without an explicit action by a named educator.
No authorship judgement
CoAcademic does not detect AI use, plagiarism, or authorship, and will not infer it.
No silent change
Rubric versions, faculty edits, and decisions are recorded rather than overwritten.
The Founding Assessment Assurance Pilot
One course team, one assignment, one confirmed rubric — audited, calibrated, reviewed against evidence, and decided by faculty. Enrolment is not open; the page explains the intended scope and what a pilot would require from both sides.