Loading...
REVISAGE Analytics: Guiding Writing Through Reflection, Visualization, and Classroom Integration
Wu, Jianmao Atmadja, Natasha
Wu, Jianmao
Atmadja, Natasha
Files
Citations
Altmetric:
Editors
Date
2026
Educational Level
ISCED Level 6 Bachelor’s or equivalent
Curriculum Area
Geographical Setting
China
Abstract
Background/Context: REVISAGE Analytics was developed at Xi’an Jiaotong-Liverpool University to strengthen academic writing in English-medium courses by pairing AI with explicit pedagogy. The platform embeds reflective learning and teacher oversight into classroom workflows, complementing— rather than replacing—human feedback.
Problem/Opportunity: Three issues motivated the innovation: (i) cognitive offloading from “auto-fix” tools that short-circuit reflection; (ii) superficial engagement with generic grammar feedback that neglects structure and diverse learner needs; and (iii) teacher overload and limited visibility into class-wide writing trends. These gaps called for AI that cultivates feedback literacy and inclusion at scale.
Methods: A pilot with 30 learners and 6 teachers used classroom deployment and post-use surveys to evaluate usability, engagement, and pedagogical value. Iterative design emphasized visual essay diagrams; rubric-aligned, strengths-based, and customizable feedback; linguistically informed analysis using academic corpora, such as Academic Word List (AWL) and Corpus of Contemporary American English (COCA); accessibility features (dyslexia-friendly mode, text-to-speech); and an AI-plus-teacher group chat. Engagement was measured via self-reported use frequency and session duration; perceived pedagogical value was elicited through teacher comments.
Findings/Outcomes: Students reported using the tool two to three times per week with sessions typically exceeding 15 minutes, and all agreed it supported writing reflection and independent revision. Learners highlighted structure visualizations and the dyslexia-friendly mode as especially helpful. Teachers valued adjustable feedback directness, classroom dashboards, and the transparency of the shared AI group chat, noting clearer visibility of rhetorical, lexical, and grammatical issues. Positive sentiment and planned adoption into three writing course sections suggest near-term reach of 200 students; participants recommended a simpler interface and clearer scoring criteria to aid integration.
Implications for Practice: REVISAGE functions as a feedback amplifier. It nudges students to diagnose and self-repair, surfaces class-level patterns for targeted instruction, and reduces the burden of individualized comments without sacrificing rigor. Built-in accessibility and visual scaffolds broaden participation for neurodivergent learners. Rubric alignment and analytics support formative assessment, curriculum coherence, and teacher-directed human-AI collaboration.
Description
Keywords (free text)
generative AI, student writing, intelligent writing assistants, feedback, reflective learning, revision, visualization, large language models