FAQ s

DeepGrade is designed to recognise minor spelling variations using intelligent pattern recognition and contextual analysis. If the intent and understanding are clearly demonstrated, students are not penalised for small spelling mistakes. However, if a school or teacher insists on specific marking criteria in these situations, the model can be configured accordingly. Persistent issues can also be flagged for teacher review and learning support.

Yes, DeepGrade flags creative or unexpected answers for teacher review. These responses are automatically highlighted with confidence scores, enabling teachers to quickly identify which ones may need manual verification.

DeepGrade uses semantic understanding and curriculum mapping to evaluate technical correctness. In such cases, it will either accept the answer if it's a valid interpretation or flag it for human review to ensure no learning opportunity is missed. If a school or teacher insists on specific marking criteria for such responses, the model can be adapted to align with those requirements.

DeepGrade provides clear, actionable feedback for each response. It highlights how well the answer met the criteria set by the school or teacher, and identifies what specifically needs improvement to reach full marks. Teachers do not need to go through the full script unless they wish to review further.

Teachers or schools can scan the papers of the entire class or cohort in one go, and any discrepancies are automatically flagged. The system performs automatic page tracking and alerts the teacher if a page is missing or if an answer appears incomplete. Missing pages can be rescanned at any time without affecting the grading process or results.

Yes, DeepGrade provides comprehensive, term-by-term performance analytics that track individual and cohort-level trends. It not only highlights gaps and declines—such as dips from formative to summative assessments—but also suggests specific areas for improvement for each student.

The platform also offers targeted recommendations for teachers, helping them guide every learner more effectively. These include instructional strategies, learning focus areas, and resources tailored to each student's needs—empowering educators to offer the right support at the right time and become even stronger mentors.

DeepGrade uses criteria-based scoring aligned with curriculum expectations. It recognises partially correct responses and allocates marks accordingly, based on well-defined rubrics or custom scoring models set by each institution. Importantly, all marking logic can be tailored to match each school's unique marking criteria and requirements—whether it's subject-specific nuance, contextual expectations, or institution-level preferences. This ensures alignment with your assessment standards while maintaining accuracy and fairness.

Absolutely. Every marked paper is stored securely and can be reviewed, annotated, or re-evaluated by authorised teachers. Concerns raised by teachers or parents are addressed through a transparent audit trail and support from our education success team. Additionally, DeepGrade offers a messaging feature linked to each question, allowing students or parents (via student login) to raise clarification requests or marking disputes. Teachers can respond directly through the system, enabling smooth, accountable communication and resolution.

Yes, DeepGrade offers a secure student portal designed to give learners full visibility into their assessments. Through their login, students can access their marked answer scripts, view detailed feedback, and understand where marks were gained or lost. The portal also provides personalised suggestions for improvement, helping each student track their academic progress over time. In addition, students can raise queries or clarification requests directly against specific questions, and—if enabled by the school—receive responses from their teachers. This creates a transparent, interactive feedback loop that encourages ownership, reflection, and continuous growth.

Yes, our system is trained to recognise and accept multiple valid solution paths for mathematical problems. We collaborate closely with schools to continuously update and localise the accepted methods, ensuring that evaluation remains fair and aligned with curriculum standards. Importantly, we can also customise this logic—restricting the use of certain methods or applying partial deductions—based on the specific marking guidelines set by each school. This allows schools to uphold their preferred teaching strategies while maintaining consistent and accurate grading.

Yes, DeepGrade supports inclusive assessment practices. If scripts are written by a scribe, they can be scanned and evaluated as normal, with additional teacher flags or comments included if necessary.

While we do not yet have formal third-party accreditation, DeepGrade actively benchmarks its scoring against globally recognised education standards and real classroom marking practices. Our in-house academic and AI experts rigorously validate all models using teacher-marked datasets and live school feedback. This ensures that our grading remains fair, transparent, and aligned with what educators expect — even as we continue to explore external partnerships and validation pathways.