I need a purpose-built AI application that can read contracts, court documents, and legal research papers, then turn that analysis into first-pass submissions and pleadings. My overriding priority is ease of use: a junior in my office should be able to drag-and-drop a PDF and receive a clean, well-structured draft without tinkering with settings or prompts. Key expectations • End-to-end workflow: upload → automated review (clause extraction, issue spotting, citation check) → editable draft formatted for court. • Modern, intuitive interface—web-based or desktop—where the heavy lifting happens behind the scenes through NLP, machine learning or large language models (Python, TensorFlow, PyTorch, or similar frameworks are fine). • Secure handling of sensitive material; on-prem or private-cloud deployment is acceptable. • Modular architecture so accuracy improvements or additional document types can be added later. Acceptance criteria 1. A working prototype that processes at least three sample files per document category and produces a submission draft in Word. 2. Response time for a 30-page document under five minutes on standard hardware. 3. No-code onboarding: users can operate the system after a ten-minute walk-through. 4. Clear hand-off package: source code, environment setup guide, and brief documentation. There’s no rigid deadline; quality and thoughtful UX matter more than speed. Let me know which existing models or libraries you would leverage and how you plan to balance transparency, data security, and future scalability.