Development of a (LLM) for Automated Will Drafting

Description

Project Overview:We are seeking an experienced developer with expertise in artificial intelligence, particularly in natural language processing (NLP), to create a custom Large Language Model (LLM). This LLM will parse client data from a structured data capture form and interpret notes provided by our admin team. Using this information, it will generate a legally compliant will based on predefined question and answer sets.
The goal is to streamline and automate the will-drafting process, ensuring accuracy, compliance, and efficiency.
Project Goals:Data Parsing: The LLM must accurately extract and interpret data from structured forms and unstructured admin notes.Predefined Logic: Use predefined question and answer sets to guide the will-drafting process, ensuring legal compliance and adherence to client instructions.Automated Will Drafting: Generate a complete, professional, and legally compliant will document tailored to the client’s unique inputs.Validation and Error Checking: Include a mechanism to flag incomplete or contradictory data for review.Scope of Work:Data Integration:
Develop functionality to parse structured data from client capture forms (e.g., JSON, CSV, or other formats).Process unstructured admin notes and combine them with form data for context.LLM Development:
Fine-tune an existing LLM (e.g., GPT, Claude) or build a custom model to interpret client inputs and map them to the predefined logic.Ensure the model respects UK legal terminology and practices for will drafting.Question and Answer Logic:
Implement predefined Q&A logic to guide the model's output.Ensure flexibility to adapt to complex client scenarios, such as multiple beneficiaries, trusts, or guardianship provisions.Will Document Generation:
Output a professionally formatted will document (Word/PDF).Include all necessary clauses, conditions, and legal provisions required for compliance in the UK.Error Handling:
Build a system to flag missing or inconsistent data.Provide actionable feedback for admin to correct inputs before generating the final document.Testing and Validation:
Thoroughly test the model to ensure it generates accurate and compliant wills.Validate the outputs against multiple scenarios to cover edge cases.Deployment:
Integrate the LLM into our existing system or provide a standalone solution with an intuitive user interface.Deliverables:Fully functional LLM tailored to parse client data and draft compliant wills.Integration with existing client data capture forms and admin note workflows.Predefined Q&A logic implementation for automated drafting.Documentation for the system, including user guide and technical specifications.