About

Project Title AI-Powered Building Survey Report Generator (Level 2 & Level 3) Overview I am an AssocRICS qualified Building Surveyor seeking a developer to build a secure web application that dramatically reduces the time needed to produce RICS-compliant Level 2 and Level 3 building survey reports. The system will combine pre-drafted paragraphs with my dictated site notes using OpenAI/LLM technology to create a complete survey report in DOCX and PDF formats. This first phase is for my internal use only (single user). Future phases will expand to include automated valuations and advanced image analysis, but Phase 1 must stand alone as a fully functional product. Primary Goals (Phase 1) Report Generation Accept dictated text notes (or uploaded text files) and property details. Allow me to select Level 2 or Level 3 survey type. Merge pre-written paragraphs with my notes using AI to populate a full RICS-compliant report. Output in DOCX (editable) and PDF (locked) formats. Reference Numbering Automatically generate a unique reference number for each new survey in the format: NABS****** (NABS + self-generating 6-digit sequential or random number). Photo Handling Upload up to 50 photos per survey. Add manual captions in Phase 1. Photos to appear within the report and in a final appendix. Data Storage & Records Securely store all generated reports, uploaded notes, and photos. Provide search/filter to retrieve past surveys. Maintain full audit trail for compliance and record-keeping. Data will also be retained so that the AI can continuously analyse previous reports to improve accuracy. Future Phases (not included in this contract, but system should allow for growth) Phase 2: AI-driven photo recognition and categorisation (e.g., identify roofs, walls, damp areas). Automatic tagging of photos to the correct report sections. Phase 3: Integration of property valuations and related outputs. Multi-user SaaS licensing to other surveyors. Core Features & Requirements (Phase 1) A. Input Property Form: Address, client details, survey level (2 or 3), date of inspection. Text/Notes: Paste or upload pre-drafted notes. Optional audio upload for automatic transcription (OpenAI Whisper or equivalent). Photos: Drag-and-drop interface supporting up to 50 images with manual captions. B. AI Processing OpenAI/LLM Integration: Classify notes into RICS report sections A0–I4. Insert pre-drafted paragraphs where applicable. Combine standard text and custom dictation into cohesive narrative. Mandatory Phrases: Automatically include key RICS compliance statements (roof inspection, certification advice, planning/building-regulations advice). Validation: Strict JSON schema enforcement for all outputs to prevent malformed data. C. Report Output Templates: Branded “Operation NABS” style with logo, colours, and fonts. Clear section codes and condition ratings (Green 1 / Amber 2 / Red 3 / NI). Forensic checklist ✔️/❌ at the end of each report. Formats: Editable DOCX using docxTemplater or python-docx. Paginated PDF using HTML→PDF (Playwright/Chromium or similar). Photo Integration: Photos inserted in relevant sections and collected in an appendix with captions. D. Data Management Database: Postgres or equivalent for storing reports, notes, photos, and metadata. Search & Retrieval: Simple dashboard to view and download past surveys. Audit Trail: Log of inputs, AI prompts/outputs, and final edits. E. Security & Compliance UK/EU hosting with GDPR compliance. Encryption of all data at rest and in transit. Single-user authentication with secure login. Technology Preferences (Open to discussion if developer proposes an equivalent modern stack) Layer Preferred Tech Front-end React + Next.js, Tailwind CSS Back-end Node.js/Next.js API or Python/FastAPI AI OpenAI GPT-4 (Chat) + Whisper for audio transcription Database Postgres (Supabase or AWS RDS) File Storage AWS S3 / Cloudflare R2 / Supabase Storage Document Generation HTML→PDF (Playwright/Chromium) + docxTemplater Deliverables Fully working web application deployed to secure hosting. Source code and documentation (installation guide, developer notes). Custom report templates (DOCX and PDF) matching my branding. User guide for daily operation. Milestones (suggested) Data input & reference number generator – property form, NABS****** ID, text/photo upload. OpenAI integration – structured JSON output, condition ratings, mandatory phrases. Report generation – branded DOCX/PDF export with photo appendix. Dashboard & storage – search and retrieval of past reports, audit logging. Security hardening – GDPR compliance, encryption, deployment. Acceptance Criteria The system successfully converts my notes + pre-drafted paragraphs into a complete Level 2 or Level 3 report with unique NABS****** reference. Reports include captions for uploaded photos and meet formatting requirements. All data is securely stored and retrievable. Output passes validation against the RICS Home Survey Standard. Intellectual Property All code, templates, prompts, and generated documentation will be the exclusive property of N A Building Surveys Ltd. Developer to sign NDA and IP assignment agreement. Summary I need a developer who can deliver a secure, AI-integrated full-stack web application that turns dictated notes and pre-written paragraphs into complete Level 2 and Level 3 building survey reports, complete with unique NABS****** reference numbers, photo uploads with captions, and compliant record keeping. This will be a Phase 1 internal-use build with a clear roadmap for Phase 2 (AI photo classification) and Phase 3 (valuations and SaaS licensing).