10172292Second Mesh: Secure AI Tools for Regulated SectorsClosedFast Start ResponseInnovate UKThe rapid advancement of Generative AI (GenAI) services has been widely welcomed by the education sector, public sector organisations, and SMEs, all of whom recognise the potential of these tools to improve productivity, streamline processes, and unlock the value of organisational knowledge. However, many of these institutions are unable to adopt mainstream AI platforms because they do not comply with sector specific data protection and compliance requirements, including UK GDPR. This gap between potential and practical adoption is where Second Mesh Ltd steps in. Our bespoke AI solutions are designed from the ground up to operate securely within each client's environment, ensuring that sensitive data remains private and fully compliant with organisational policies and the regulations which govern their sectors. By aligning with the specific needs and standards of regulated sectors, our bespoke AI services make GenAI services trusted and accessible. Second Mesh is building secure, tailored Generative AI services that help UK education providers, public sector organisations, and SMEs make better use of their own data. Our platform allows organisations to create private AI chatbots that are grounded in their internal document libraries and systems, delivering relevant, trusted answers while keeping all data processing fully under their control. We use open-source, open-weight large language models to ensure transparency and customisation, and our modular AI toolset makes it easy to deploy safe, cost-effective AI solutions at speed. This project will support further development of our toolsets so that they can be used confidently in regulated environments like schools, colleges, local authorities, and the SME that serve them. Second Mesh wishes to demonstrate that organisations such as educational institutions and other organisations who operate within high stakes and highly regulated sectors can make use of hyper local language models, and to do so in a manner that is affordable, private, secure and which have greater transparency and governance arrangements than the GenAI chatbot services which currently dominate the AI market place. During the funded period, we will undertake work on two work packages which will enable us to develop, deploy and scale the use of local large language models that can be used within education institutions and other regulated sectors. The first work package entails the development our AI Toolkit which will be compliant with AI safety standards. The second work package will see the advent of scalable local AI infrastructure and model deployment.