How to build a Writing Feedback Assistant
This agent reviews drafts for clarity, tone, structure, and correctness, then suggests targeted edits with examples.
Challenge
Writers and students lose time on manual edits, inconsistent feedback, and back-and-forth reviews. Style and tone drift across authors, while common issues like passive voice, jargon, and weak structure slip through. Editors get overloaded, slowing publication and creating bottlenecks. The result is uneven quality, missed deadlines, and writing that underperforms.
Industry
Education
Operations
SaaS
Department
HR
Content Creation
Integrations

OpenAI
TL;DR
What it does:
Automates the process of reviewing student essays by:
Accepting a student’s name and essay upload
Accepting assignment guidelines upload
Using AI to generate feedback based on the essay and guidelines
Formatting the feedback for the student
Logging the feedback and file references to a Google Sheet for teacher review
Who it’s for:
Teachers and educators who want to streamline essay grading and feedback
Students submitting essays for assignments
School administrators seeking to centralize and track student feedback
Time to value:
Immediate: Once set up, teachers and students can use it right away
No manual review or formatting needed—AI and automation handle everything
Output:
A well-formatted feedback report for the student, shown instantly
A log entry in a Google Sheet with feedback, essay file link, and guidelines file link for teacher records
Common Pain Points of Providing Feedback
Manual essay review is time-consuming and inconsistent
Teachers struggle to provide detailed, criteria-based feedback for every student
Tracking which feedback was given to which student can be messy
Students may not understand vague or generic feedback
Administrative overhead in logging and organizing feedback for future reference
What the Agent Delivers
Automated, criteria-based feedback: AI reviews essays against provided guidelines, ensuring fairness and clarity
Instant, formatted student reports: Students receive clear, structured feedback as soon as they submit
Centralized teacher records: All feedback and file references are logged in a Google Sheet for easy review and auditing
Reduced teacher workload: Frees up time for teachers to focus on instruction, not paperwork
Consistency: Every student is evaluated against the same rubric/guidelines
Want to see this agent in action? Check out the video tutorial below.
Step-by-Step Build (StackAI Nodes)
1) Student Name (Input Node)
What it does:
Collects the student’s name as text input.
This is a simple text field; connect it to downstream nodes if you want to personalize feedback or log the student’s name.
Goal:
Identify which student the essay and feedback belong to.
2) Student Essay (Files Node)
What it does:
Allows the user to upload the student’s essay file (supports various formats).
Extracts the text for AI processing.
Set as user-exposed so students/teachers can upload files directly.
Goal:
Provide the essay content for review.
3) Assignment Guidelines (Files Node)
What it does:
Allows upload of the assignment’s grading criteria or rubric.
Extracts the text for AI reference.
Also user-exposed for easy upload.
Goal:
Give the AI a clear standard to compare the essay against.

4) OpenAI (LLM Node)
What it does:
Uses GPT-4o-mini to analyze the essay in the context of the guidelines.
Generates detailed, criteria-based feedback.
The system prompt is tailored for Romeo and Juliet assignments, but can be adapted for other topics.
Receives essay and guidelines as input.
Goal:
Automate the feedback process with high-quality, rubric-aligned comments.

Instructions
Prompt
5) Feedback Template (Template Node)
What it does:
Formats the AI’s feedback into a clear, student-friendly markdown report.
You can customize the template for your school’s branding or feedback style.
Goal:
Ensure feedback is easy to read and professional.
6) Write Sheet (Action Node)
What it does:
Logs the feedback, essay file link, and guidelines file link to a Google Sheet for teacher review.
Uses a pre-configured Google Workspace connection and spreadsheet.
To log data, you must specify:
The Google Sheet to write to (spreadsheet_id)
The specific sheet/tab (sheet_name)
The data to write (as a JSON object or plain text)
The data field can include key-value pairs like:
{"Essay Feedback": "{llm-0}", "Essay File": "{doc-0.file_urls}", "Guidelines File": "{doc-1.file_urls}"}
Goal:
Centralize all feedback and submissions for easy tracking and auditing.

7) Output (Output Node)
What it does:
Displays the formatted feedback to the student immediately after processing.
Connects directly from the template node for clean output.
Goal:
Provide instant, actionable feedback to the student.
