I want an end-to-end Claude-powered agent that brings all of my buy-side research workflows under one roof. The agent must • Automatically pull PowerPoints, “Concall Transcripts”, annual reports from BSE or screener.in, and broker research (updates, downgrades, initiating-coverage notes) straight out of my one-year email archive • Parse each file, detect the company it belongs to, and save it in a dedicated Drive folder for that company—no mix-ups • Pass every saved document into NotebookLM and spin up a fresh notebook per company, ready for question-answering inside NotebookLM • Sit inside Microsoft Teams as a bot so I can type, for example, “Create notebook for HDFCBANK” or “What did the latest downgrade say about margins?” and get a reply on the spot • Turn any notebook into a podcast episode, then push the audio straight to my private YouTube channel Also automatically update notebooks whenever you there are any new documents on our emails or new results for the company Core expectations – Zero-touch scraping from BSE (public filings) plus IMAP/Graph harvesting for the broker emails – Robust company-name attribution logic so documents land in the right Drive folder every time – Reliable Claude calls for summarisation, Q&A and podcast script generation – Seamless hand-off to NotebookLM and YouTube APIs – Teams bot built via Microsoft Graph with slash-style commands for create, ask, podcast, upload Acceptance criteria 1. I say “/create ITC” in Teams → a Drive folder, NotebookLM notebook and confirmation message appear in under two minutes. 2. I forward a broker report to the tracked mailbox → it reaches the correct folder and notebook automatically. 3. “/podcast ITC” produces an audio file on the private YouTube channel with proper title, description and chapter markers. 4. Source code, environment files and a short README are delivered so I can redeploy internally. If you have solid Python-API chops, experience with Claude or similar LLM orchestration, and can wire everything together securely, let’s talk.