CAPABILITY EVIDENCE
The assessment model for work students can explain
AI detectors try to answer “who wrote this?” Capability evidence answers a better question: “can the student explain what they submitted?”
WHAT IS CAPABILITY EVIDENCE?
Structured proof that a student can explain the work they submitted.
A submitted essay, codebase, or presentation tells you what was produced. It does not tell you whether the student understands it. Capability evidence closes that gap.
After submitting their work, students complete a short adaptive follow-up: targeted questions generated from their specific submission and mapped to your rubric. Their responses are transcribed, scored, and linked back to the artifact so you can review the evidence before making a grading or integrity decision.
The result is a reviewable record that connects four things: the artifact (what was submitted), the explanation (what the student said about it), the rubric alignment (how responses map to your criteria), and the instructor decision (your final judgment).
THE PROBLEM
AI detection cannot prove learning
AI detectors analyze writing style. They return a probability that text was machine-generated. But a low AI score does not mean the student understands the material. A high score does not mean they cheated. The output is a statistical guess about authorship, not evidence of learning.
Educators need something different: a way to ask “does this student understand what they submitted?” and get a structured, reviewable answer. That is what capability evidence provides.
THE EVIDENCE LOOP
From submitted work to explainable evidence
Student submits work
The student uploads their artifact: an essay, code, slide deck, lab report, or any other format. This is the starting point, not the finish line.
Adaptive follow-up begins
Pruuva generates questions tied to the specific submission and your rubric. The student explains their work through text, audio, or video responses.
Evidence is structured
Responses are transcribed, scored against rubric dimensions, and linked back to specific passages in the submission. Gaps and strengths surface automatically.
Instructor reviews and decides
You see the artifact and evidence side by side. Agree with the AI assessment, adjust scores, add annotations, and make the final call. Your judgment is always authoritative.
WORKS WITH ANY ARTIFACT
Evidence for every type of student work
Capability evidence is not limited to essays. Pruuva generates follow-up questions from whatever the student submits.
Essays and research papers
Follow-ups probe methodology, source selection, and whether the student can explain the argument they constructed.
Code and technical projects
Questions target design decisions, edge cases, and implementation choices. Students explain the code they wrote, not just the output it produced.
Presentations
Students present their work live via screen share, then answer targeted questions about their data, reasoning, and design choices.
Lab reports and fieldwork
The defense explores experimental design, data interpretation, and whether the student can connect results to the underlying theory.
Group projects
Each team member gets their own follow-up. Individual understanding is assessed even when the deliverable was collaborative.
Creative and design work
Questions explore influences, process decisions, and the reasoning behind creative choices that are hard to evaluate from the output alone.
WHO IT HELPS
Capability evidence helps institutions move from trust by assumption to trust by review
For faculty
Grade with structured evidence instead of intuition. When you suspect a student didn't do the work, you have a fair, documented way to verify. When they did, you have the evidence to support the grade.
For departments
Review patterns across sections and semesters. Understand whether assessments are producing evidence of learning, and identify where capability gaps cluster.
For institutions
Replace ad-hoc integrity workflows with a reviewable evidence standard. Support accreditation requirements, reduce appeals disputes, and give policy teams something defensible to point to.
IMPLEMENTATION
Add capability evidence without redesigning every course
Students submit their work exactly as they do now: PDF, DOCX, code, or pasted text. Pruuva attaches to your existing assignments and generates follow-up questions from each student's specific submission.
You configure the rubric dimensions, choose the response mode (text, audio, or video), set the deadline, and publish. Students receive a link, complete the follow-up on their own schedule, and you receive a structured evidence report.
Most educators start with a single assignment in one course. A typical pilot runs for one semester, covers 1 to 3 sections, and produces enough data to evaluate whether capability evidence improves grading confidence and reduces integrity disputes.
FREE RESOURCE
Capability Evidence Checklist for AI-Era Assessment
A practical guide for evaluating whether your current assessment process produces reviewable evidence of student understanding. Covers rubric alignment, follow-up design, evidence standards, and implementation steps.
TRUST AND ADOPTION
Human judgment, privacy, accessibility, and data processing
Instructor authority
Pruuva surfaces evidence and suggestions. The instructor always makes the final decision on grades and integrity outcomes. The AI assists; it does not adjudicate.
Student fairness
Students choose how to respond: text, audio, or video. Extended time and accommodation settings are built in. The process is a structured conversation, not a surveillance tool.
Data processing
Student submissions and responses are processed to generate evidence and are not retained beyond the instructor's review period. See the AI processing and trust page for full data-handling details.
Institutional control
Institutions configure retention policies, access controls, and export formats. Evidence belongs to the institution and the instructor, not to Pruuva.
COMPARE YOUR CURRENT APPROACH
See how evidence-based assessment differs from the tools you may already use
AI DETECTION ALTERNATIVE
When detector scores are not enough
Compare Pruuva with AI detection workflows when your review needs student explanations, rubric context, and evidence a teacher can inspect.
Compare AI detectionTURNITIN ALTERNATIVE
When similarity is only one signal
See how capability evidence changes the review from text matching to demonstrated understanding of submitted work.
Compare TurnitinFULL COMPARISON GUIDE
Choose the right integrity path
Review all comparisons if you are deciding between detectors, proctoring tools, lockdown browsers, and evidence-based assessment.
View all comparisonsCOMMON QUESTIONS
Questions about capability evidence
What is capability evidence?
Capability evidence is a structured record showing that a student can explain the work they submitted. It connects the original artifact, the student's follow-up responses, rubric-level findings, and the instructor's final decision into a single reviewable package.
How is this different from AI detection?
AI detectors scan text for statistical patterns and return a probability score. Capability evidence asks the student to explain their own work, then maps those explanations to your rubric. The output is structured evidence of understanding, not a binary flag.
Does capability evidence replace grading?
No. It gives you better information to grade with. You still assign the grade, make integrity decisions, and control the rubric. Pruuva surfaces the evidence. You make the call.
What if a student has test anxiety or accessibility needs?
Pruuva supports text, audio, and video response modes. Students go at their own pace, can pause and resume, and have extended time accommodations available. It is a structured check-in, not a high-stakes timed exam.
Can I use this with my existing assignments?
Yes. Students submit their work as usual. Pruuva generates follow-up questions from their specific submission and your rubric. No course redesign required.
How long does it take for students to complete?
Most follow-up sessions take 10 to 20 minutes. The length adapts based on the complexity of the submission and the rubric dimensions you configure.
Is the evidence admissible in academic integrity proceedings?
Capability evidence provides a timestamped, structured record with student responses, rubric scores, and instructor notes. Institutions can use this as part of their integrity review process. The record is designed to be clearer and more defensible than a detector score alone.
What happens to student data?
Pruuva processes data to generate evidence and does not retain submissions or responses beyond what is needed for the instructor's review. See the AI processing and trust page for full details on data handling.
RELATED RESOURCES
Pilot capability evidence in one course
Start with a single assignment. See the evidence it produces. Decide whether to expand. No institutional commitment required to begin.