UPSC candidate credits AI tools for securing rank of 107

Mohammad Owais Bhat achieved an All India Rank of 107 in the Combined Defence Services examination conducted by the Union Public Service Commission. This result marks his successful third attempt at the competitive exam.The candidate utilised generative AI for syllabus-based trend analysis and question format prediction, achieving a top percentile rank.

The figures

The source material provides specific metrics regarding Bhat’s academic achievement and preparation strategy. His preparation spanned more than four years, with some accounts describing it as four-and-a-half years. He subscribed to an artificial intelligence service for ₹399. By analysing question formats with AI, he claimed to anticipate between 20 and 40 questions beforehand.

Behind the figures

Bhat described his methodology as a targeted approach to managing the vast syllabus. He advised aspirants to study the syllabus thoroughly first, comparing unstructured preparation to an endless ocean that can be narrowed into manageable ponds with focus. He leveraged artificial intelligence tools, specifically ChatGPT and other bots, to support this strategy. He fed the syllabus into the AI tool and prompted it to predict ongoing trends. This process helped him identify patterns and anticipate question formats.

What changes, what does not

Bhat’s success highlights a shift in preparation methods for competitive examinations in India. He noted that effective preparation is possible without expensive coaching centres, recommending the use of YouTube lectures instead. His achievement brought recognition to his family and neighbourhood in the Dangerpora area. He cited Brigadier S.S. Shekhawat, a decorated Army officer, as a major source of inspiration. The physical demands of the selection process remain a significant factor alongside academic preparation.

What builders can do with it

Developers can observe how users are applying large language models for pattern recognition in structured testing environments. Bhat’s use case demonstrates feeding a syllabus into an AI tool to predict trends and identify question formats. This suggests that builders can design interfaces that allow users to upload curriculum data and receive trend analysis. The documented limit is the user’s ability to prompt the AI effectively to anticipate specific questions. Builders can focus on creating tools that help users narrow down vast information sets into manageable insights, mirroring Bhat’s strategy of converting an "endless ocean" of content into specific, actionable study points.