I want to turn every spoken doctor’s note into a full workflow that finishes with a prescription on the patient’s WhatsApp and clean data inside Power BI. In practice, that means: • The system must take an audio stream or file, transcribe it accurately, run the text through an OpenAI (or similar LLM) prompt that structures it into a standard EMR entry, and auto-generate a prescription. • Once the prescription is produced, the platform should send it to the patient through WhatsApp (API-based, e.g., Twilio or Meta’s Cloud API). • Both the finalized patient record and the prescription data need to flow—ideally near real-time—into a Power BI dataset so the operations team can track and optimise clinic performance. I will supply sample audio, note templates, and the schema for the records table that feeds Power BI. Code needs to be written in Python, with clear, commented functions that let us swap out speech-to-text or LLM providers later. If you already have quick POCs using Whisper, Azure Speech, or similar engines, let me see them. Acceptance criteria 1. Doctor speaks, JSON record appears in under 30 sec with ≥95 % word accuracy in key medical fields. 2. WhatsApp message arrives automatically with the prescription PDF and simple plain-text summary. 3. Power BI refresh picks up new rows without manual intervention; record count in BI matches the database. 4. Codebase is dockerised, documented, and ready for hand-off or continued collaboration. If this sounds like your kind of build, share one live example of an AI pipeline you created and any relevant repos or demos.