All shot-based learning techniques are very much useful for train the models just with a small number of datasets. The models are trained by large datasets always producing a better result but it is very challenging, tough or expensive task to collecting the large datasets. Therefore, with the help of prior knowledge and small datasets, all these learning approaches providing benefits to different domains like Natural Language Processing (NLP), robotics and image recognitions. All these learning approaches faced a lot of challenges because of the complexity of their learning process with a small dataset and it leads to in accurate results and it is also toughest job to generalize unseen data. Hence, this research work aims to educate the end users with different query formats while their interaction with ChatGPT and also enhancing the model’s accuracy by adopting techniques like transformers and fine-tuning. Each shot-based approach provides different types of levels of accuracy that is totally depending upon user’s request, but N-shot learning approach always produced better result consistently when compared to other learning approaches. By utilizing well-structured query formats helps to the end users to get more effective and accurate responses from AI models like ChatGPT with N-shot learning approach.