TDR Sahayak, AI Refund Guidance Hackathon

video1.0<iframe src="https://www.loom.com/embed/9da764cdb88f4f7283a36e1cbaf23839" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/os/loom-media-content/loom/fae5d4d5-6fcb-4aa9-aeb0-c465ce37f7dc/4fcf76cf-215c-4e7c-af03-0b940566d3c2/sessions-thumbnails-9da764cdb88f4f7283a36e1cbaf23839-3deb855a6d642435.gif/binary?authToken=eyJraWQiOiJtaWNyb3MtZ3JvdXAvdGRwLW9zLWFyY2hldHlwZS9icjZlYTZwZmhrdm5tcWJsIiwiYWxnIjoiUlMyNTYifQ.eyJicyI6Imxvb20tbWVkaWEtY29udGVudCIsImF1ZCI6InRkcC1vcy1hcmNoZXR5cGUiLCJuYmYiOjE3OTM1NTYyOTEsInRpIjoiZmEyZDhiOWMtZjFiNi00YjUwLWJhMmYtYzU2MDVhNGFmOGE5IiwiYWRpIjoibG9vbS9sb29tIiwiaXNzIjoibWljcm9zLWdyb3VwL3RkcC1vcy1hcmNoZXR5cGUiLCJwaWQiOiI0YzllYzg5YS02MGNiLTRhZWItOGNhMS00MTE4YTg2NDA0NmQiLCJ3YSI6ImZhbHNlIiwib2lkIjoic2Vzc2lvbnMtdGh1bWJuYWlscy05ZGE3NjRjZGI4OGY0ZjcyODNhMzZlMWNiYWYyMzgzOS0zZGViODU1YTZkNjQyNDM1LmdpZiIsImV4cCI6MTc5MzU1Njg5MSwiaWF0IjoxNzkzNTU2MjkxLCJqdGkiOiJjNWE4YTA0MC1hZjdiLTRmODAtOWNiOS05ZWQ4NmRkOGY5NjgifQ.tm4uZDTxFIbxApHbE_0SNoohISANNzqOKQ0PpyZ2nZqshKaoJz1hg5lYEUJR0YHsolVRvyM3wrzQTm6ODdKfcaXIHE3IO3KWYTG-nidef15ElHdji4JyE9LVLqHWklixCrGQBybK3BZ5TWxWbueleYfv0-kRN2Bzyxc8r9e7gf1DEX0fWAYcqNfp0ZkYO8bjUteb4Sn9WTaJkJbE6v6zuUFPWrTo1elG018V3Rbq3wolWHcC0wvqpdfAkbKOQuVDP1_03kPBVEggOhbv4m0j1UKjcDbaIgCCv8MXpWwccAcw3t_eSiLLz7bPpGsjzaRH8rxteO2WLNh-JG8ioC-DQw&Expires=1793556892&Signature=MfenrAPDpuD50rCZT8XgfN-luXbgUyb7utwAa95yVg7PArsT9BvP5Tx66FHaEyK-imgTRm69itQ9HoTdGx5kLxGrKjc3ujP-1WexaZrKR9H%7EIvfS4DzkfoQ2Ut8HQPakuDe4J8yb-4vIUFC255I3LhEe4YzEDKiiijZ6vZJ1rSwUh825bVincBEBOcJ66VzM341YT1p8loFcCSqN87M5B1f8t0umlgmPAj4vQPq4djTXyV6eFwfQcGa7iTENJnSKLMl3UFKP4-ElHSfN1DoV5O9ye7j9O19w9VNqCiTfSle%7Eg4e4Io5vnn161AncCJSf6FMrWPtCKEsjrCCnu9IdMg__&Key-Pair-Id=KQOSYIR44AIC0152.341TDR Sahayak, AI Refund Guidance HackathonThis Loom presents TDR Sahayak, a hackathon tool that helps citizens handle journey disruptions and claim refunds. The presenter demonstrates the UI where a user can type a query like “train delayed and I did not travel,” and the system responds with a risk level, what to check before filing, and guidance to prepare a mock claim with editable details. They explain the frontend was built using React with Tailwind and TypeScript for a smooth interface. On the backend, the user query is processed by a “GPT-mini” module that converts natural language to queries, while the claim part uses a separate component or an inbuilt parser to convert scripts when GPT is not used.