Agent AID FINAL

The five characters are in different cities all over the world. They are what a future architect looks like. One of them is in China, one of them is in Myanmar, and one of them is in Indonesia. And the other two years in Singapore. They are raising bots who help them to realize their architectural vision. These bots (Agent A, Agent I & Agent D) help them with different tasks in an architectural process. AGENT A - an innovative AI-driven architectural assistant designed to automate the creation of building models within Autodesk Revit. By addressing the inefficiencies of manual drafting and the complexity of Singapore’s regulatory codes, the tool aims to reduce BIM generation time from days to minutes. The system utilises Retrieval Augmented Generation (RAG) and agentic workflows to ensure that every design iteration is inherently compliant with local standards. A focused Proof of Concept targets the automated design of multi-storey car parks to demonstrate significant savings in man-hours and costs. Ultimately, the solution seeks to shift the architect's role from repetitive drafting to high-level design oversight by integrating intelligent automation directly into the existing professional workspace. AGENT i, an innovative AI conversational chatbot designed to bridge the gap between complex BIM software and project stakeholders. By leveraging GenAI and automated workflows, the tool allows users to query architectural data using natural language without requiring technical knowledge of Revit. The project aims to eliminate data silos and manual extraction pains by synchronising live information with Notion and an interactive project dashboard. This solution promises significant operational efficiency, including a 90% faster response time for queries and a substantial reduction in report preparation effort. Ultimately, the framework employs a standardised data pipeline to ensure a single, accessible source of truth for design and management teams. AGENT D, an AI-driven architectural data manager designed to automate the preparation and validation of BIM submission data. Current manual processes are plagued by repetitive tasks and high error risks, often taking months to complete for complex projects like Singapore’s CORENET X mandates. By implementing an agentic workflow, the project aims to reduce data processing times from two months to just two days, ensuring compliance with contractual and regulatory standards. The solution leverages generative AI and Python-based tools to shift validation tasks earlier in the design phase, providing real-time asset data for facilities management. Ultimately, this digital transformation replaces rigid rule-based checks with autonomous reasoning, significantly lowering costs and resource demands for large-scale construction projects.