Cheribim Ltd, in partnership with Sheffield Hallam University, are creating next generation "chat-to" capability to enable historic building owners and managers to have conversations with best practice guidance. This is to help anyone navigate the sometime complicated and confusing, funding, permissions, retrofit and adaptation advice for net zero, and consultants' recommendations by making an expert AI Assistant, able to answer specific questions about their buildings, and put advice in the context of their circumstances and level of understanding.

Cheribim Ltd, in collaboration with Sheffield Hallam University (SHU), are exploring how Large Language models (LLMs), fine-tuned with retrieval-based approaches can be used to enhance the condition inspection process at historic buildings, encouraging appropriate conservation planning and activity, including that aimed at transition to net zero at local, regional and national level. This project seeks to maximise the benefits of retrieval-based fine tuning, including its flexibility, efficiency, cost effectiveness, and the potential for improved accuracy. We are developing an AI-enabled report analyser capable of responding to a variety of enquiries related to the condition of historic buildings using current and legacy inspection data using a dataset derived from hundreds of Quinquennial Inspection Reports related to historic church buildings. Aimed at the cultural heritage and built environment sectors, the analyser is specifically intended for use by building managers and central officers involved in the day-to-day maintenance of these important cultural places. We are working closely with custodians and managers to identify the most important of these enquiries and define clear 'red lines' - where users are advised to consult conservation experts such as their architect or surveyor. The report analyser is intended to support appropriate conservation activity based on expert assessment, ensuring the protection of historic fabric. Key features will include: * A natural language interface, capable of responding to specific enquiries in an accessible, clear way, like Chat GPT and Microsoft CoPilot * The ability to identify building elements (e.g. roof) and assign related condition observations and recommendations * Insights based on single reports, multiple reports associated with one building, and multiple reports across buildings

Cheribim Ltd, in collaboration with Sheffield Hallam University, are exploring how natural language processing AI, like that used in GPT4, could be used to enhance and speed up condition inspections at historic buildings. Conservation architects and surveyors often spend up 75+% of the time it takes to deliver an inspection report to a building owner, just writing up and formatting reports. There's huge potential to cut this time completely, freeing them to use their skill and experience on higher impact work like specifying repairs and designs.

Cheribim Ltd and The Diocese of Sheffield (Church of England in South Yorkshire) are working in partnership to develop a novel AI-enabled fabric performance assessment technique specifically for historic buildings. This will use a multitude of measurements and indicators to provide accurate ratings for best-case thermal performance, real-time thermal performance, and projected efficacy of retrofit solutions. In addition to having applicability to a variety of historic public buildings such as town and city halls, libraries and museums, future development of the assessment tool could see its application to the pre-1919 domestic building stock, vastly improving the accuracy EPC ratings of over 4.7million houses and informing owners on the safest and most effective actions to improve their energy efficiency and reduce carbon emissions associated with heating, lighting (and living in!) their homes. Churches host activities which deliver over £50 billion-worth of social value to our communities, including food-, warm- and cool-banks. However this is threatened by rising energy and upkeep costs, and the challenges and uncertainties of improving and adapting their buildings. This project will help assure their path to net zero, safeguarding against inappropriate and damaging retrofit and helping them select the highest impact measures to reduce their energy consumption and associated carbon footprint.

Cultural Heritage BIM with IoT for energy efficiency.