Can AI Grade Handwritten Exams? Yes: Here's How DASES Does It
Yes, AI can accurately grade handwritten descriptive exams. DASES achieves 98% rubric accuracy by combining handwriting recognition with criterion-based evaluation. Learn how it works.
Step-by-step quickstart guide for DASES. Create an account, build your first question paper, set rubrics, upload answer sheets, review AI scores, and publish student reports — in 15 minutes.
Setting up your first exam on DASES takes approximately 15 minutes. The process: create a free account, create a new paper using the paper builder (upload your question paper PDF or type questions), enter model answers for each descriptive question and review the AI-generated rubric, scan your answer booklets and upload the PDF batch, wait 15-25 minutes for AI processing, review the score dashboard and approve or adjust results, then click Publish to release branded PDF reports to the student portal. No technical expertise is required.
Before creating your first exam in DASES, gather three items. First, your question paper — either a PDF of the printed paper or the questions typed out. Second, your model answers — the ideal responses for each descriptive question. These can be written directly in the platform or pasted from an existing document. Third, your scanned answer sheets — after the exam has been administered, the physical booklets need to be scanned as PDF files. Any office scanner or multifunction printer set to at least 300 DPI will work. The answer sheets can be scanned as individual files per student or as one combined multi-student PDF; DASES handles both formats.
Visit dases.in and click "Start Free Trial." Enter your name, institutional email address, and institution name. DASES will set up your institution profile — including your logo upload and institutional branding preferences, which appear on all student PDF reports. If your institution already has a DASES account, your administrator will invite you via email to join the existing institution workspace. Once inside, the dashboard presents three main sections: Papers (your question paper library), Batches (in-progress and completed grading batches), and Reports (published student results and analytics).
Click "New Paper" and give the paper a name (e.g., "Microeconomics — Unit 3 Internal Assessment"). Select the exam type — Internal Assessment, End-Semester, Class Test — and enter the total marks and duration. You can upload your question paper PDF and DASES will extract the questions automatically, or you can type or paste them directly. For each question, specify the marks and question type. Once questions are entered, proceed to the next step: entering model answers.
For each descriptive question, type or paste the ideal answer in the model answer field. DASES analyzes this text and generates a detailed rubric automatically — breaking the expected response into specific criteria with suggested mark weights. Review each criterion: adjust weights if needed, add or remove criteria, and add "alternative acceptable answers" for questions with multiple valid approaches. This rubric setup is the most intellectually engaged step in the process; allocate 5-10 minutes for thorough review. A well-constructed rubric produces the most accurate AI grading output.
Navigate to "New Batch" under your paper. Import your student roster (CSV upload with student names, roll numbers, and email addresses) or enter students manually for small classes. Upload your scanned answer sheet PDF or PDFs. DASES identifies student booklet boundaries automatically in multi-student scans. Click "Start Grading." The AI pipeline begins processing immediately — segmenting pages, reading handwriting, mapping answers to questions, and scoring against your rubric. For a 60-student batch, processing completes in approximately 20-25 minutes. You will receive a notification when the batch is ready for review.
The review dashboard presents all AI-graded scores across your student batch. Answers flagged as low-confidence are highlighted in orange — click any of these to view the student's original handwriting alongside the AI's evaluation and score. Make adjustments where needed using the inline override tool. When satisfied with the results, click "Publish Results." DASES instantly generates branded PDF reports for every student, updates the student portal with their individual results and feedback, and creates an exportable result register for your records. Students with registered email addresses receive an automatic notification that their results are available.
The 15-minute setup estimate applies to a new paper created from scratch. On your second exam, the process is significantly faster: you can duplicate an existing paper, modify questions as needed, and reuse rubrics from the library with minor adjustments. Most faculty who use DASES for an entire semester find that internal assessment setup time drops to under 5 minutes per new exam after the first two cycles. The paper library and rubric bank become more valuable over time, transforming each subsequent exam setup into a review-and-adjust task rather than a creation task.