10105309Leveraging Language Models for Bottom up Construction ExcellenceClosedCollaborative R&DInnovate UKConstruction?output in the?UK?is more than £116 billion per annum and contributes 7% of GDP (ONS, 2021) but remains the least productive industry in the UK economy, at more than 20% points below the average output per hour for the whole economy. The project's core objective is to address these productivity challenges by harnessing the power of AI and Machine Learning. By 2030, it is anticipated that AI will play a substantial role in project management tasks, making it essential for the construction industry to leverage AI to enhance efficiency. However, the construction sector currently lacks the high-quality datasets and Language Model Models (LLMs) necessary for AI's productivity-boosting potential. This feasibility's focus is how to leverage our existing human centric and high-quality construction specific dataset to create an industry-specific LLM. The LLM will be built using our existing proprietary application (SymTerra), which serves as a dynamic data pipeline, delivering a consistent stream of updated project data to the LLM. This ensures that the model evolves in tandem with real-world operations, maintaining its relevance and accuracy over time. Our bottom-up approach, specifically focused on construction, bridges the critical gap by providing the human context up through the entire supply chain, a dimension other solutions overlook. Unlike OpenAI, this LLM is specific to the industry and built and trained from the human knowledge at the coalface of project delivery.