Structured feedback based on specific queries is crucial for users and service providers in various sectors, such as education, industry, entertainment, and healthcare. This process enables all stakeholders to obtain specific and direct feedback, helping them gauge their interaction with the resources/material provided and improving overall human-computer interactions. Incorporating general and specific feedback mechanisms, especially in e-learning, must be strengthened to enhance student and teacher satisfaction while interacting with e-learning material, e.g., lecture videos. The proposed work explores deep learning language models that can take in narrative reports (student/teacher feedback reports) built using user feedback and generate responses to specific questions posed by students/teachers. This process supports the requirements of students and lecturers who want to reflect on particular aspects of their learning/delivery. Usability studies reported that a large percentage of the responses (80% and 90% for single and group reports, respectively) generated during the experimental evaluation were in line with the questions posed, suggesting that the proposed pipeline performed well in response generation. Automating responses by synthesizing narrative reports by utilising a language model has the potential to provide insights into student learning affect. The proposed model is limited by the narrative reports produced by the previous models in the cascade. When incorporated, other modalities linked to learning can improve outcomes and result in a robust system.