AI³ Delivery Partners Project

AI Lab: Rapid, time limited exploration and assessment of AI use cases and experiments. The Department for Business and Trade (DBT) Digital, Data and Technology (DDaT) team is implementing the ‘AI³’ model to identify, test, develop, and evaluate use cases for artificial intelligence across departmental functions. This model comprises three development phases: • AI Lab: Rapid, time limited exploration and assessment of AI use cases and experiments. • AI Factory: Development and operationalisation of experiments into production ready tools and components • AI Operations: Maintenance, scaling, and operationalisation of AI solutions This lot is focused on the AI Lab phase, which is key to the success of the wider AI³ model. Objectives: The Supplier must provide specialist resources to: • Identify high value AI use cases through structured engagement with HMG stakeholders • Assess the feasibility and potential impact of these use cases • Rapidly prototype and test working experiments, with clear identification of success measures and experiment scope • Ensure ethical, secure, and compliant use of data throughout the process • Support decision making for onward investment and scaling Scope: The Supplier, through provision of its specialists to the Buyer, will be expected to deliver the following through agreed SOWs that will refine the specific deliverables: Use Case Identification: o Facilitate ideation workshops and horizon scanning activities o Engage with the Buyer’s (and potentially wider HMG’s) internal teams to uncover operational challenges and innovation opportunities o Reuse and adapt existing AI solutions from across government and industry Technical Feasibility Assessment o Apply data science and machine learning engineering expertise to evaluate proposed use cases o Recommend appropriate technical approaches and architectures o Collaborate with the Buyer’s teams to ensure alignment with existing infrastructure and standards Data Ethics and Governance o Ensure all AI use cases comply with ethical standards and data governance policies o Advise on responsible AI practices, including bias mitigation and transparency o Support GDPR compliance and data protection assessments Prototyping and Experimentation o Rapidly develop prototypes using standard design patterns for government services o Design and execute experiments to validate concepts o Document outcomes and provide recommendations for further development or discontinuation Collaboration and Integration o Work alongside civil servants and embedded teams o Contribute to lessons learned and continuous improvement exercises Please refer to the ITP (Appendix B) for more information on the required skills and roles needed for this lot.

AI Factory: Development and operationalisation of experiments into production ready tools and components. The Department for Business and Trade (DBT) Digital, Data and Technology (DDaT) team is implementing the ‘AI³’ model to identify, test, develop, and evaluate use cases for artificial intelligence across departmental functions. This model comprises three development phases: • AI Lab: Rapid, time limited exploration and assessment of AI use cases and experiments • AI Factory: Development and operationalisation of proofs of concept into production ready tools and components • AI Operations: Maintenance, scaling, and operationalisation of AI solutions This lot is focused on the AI Factory phase, building on the AI Lab. Objectives/Scope: The Supplier must provide specialist resources to develop and operationalise AI into services, tools and components, building on successful experiments from the AI Lab: This will include: • Service and technical architecture design informed by data expertise to translate proofs of concept into robust, user-facing tools and services • Innovation in delivery and application of AI technology, going beyond standard tools and technologies to deliver measurable applications of AI. • Assessing, designing and optimising data science and ML engineering approaches to ensure that technical AI proofs of concept perform in a real-world environment • Partnering with other digital teams within the Buyer’s DDaT team to co-deliver new AI features and components within existing digital and data services • Product and delivery expertise informed by data to ensure rapid, tightly scoped delivery of services within a few months • User-centred design expertise to ensure that products and tools are able to meet user needs, and to identify opportunities for re-use or wider roll-out of products developed • Helping to build skills around operationalisation of AI among civil servants, including core technical and infrastructure skills • Evaluation and performance analysis expertise to identify meaningful metrics for assessing the performance of AI products being delivered and enabling continuous improvement This is not an exhaustive list, and specific deliverables will be defined, refined and agreed with the commission of each SOW. Please refer to the ITP (Appendix B) for more information on the required skills and roles needed for this lot.

AI Operations: Maintenance, scaling, and operationalisation of AI Solutions. The Department for Business and Trade (DBT) Digital, Data and Technology (DDaT) team is implementing the ‘AI³’ model to identify, test, develop, and evaluate use cases for artificial intelligence across departmental functions. This model comprises three development phases: • AI Lab: Rapid, time limited exploration and assessment of AI use cases and experiments • AI Factory: Development and operationalisation of experiments into production ready tools and components • AI Operations: Maintenance, scaling, and operationalisation of AI solutions This lot is focused on the AI Operations phase, building on the AI Factory . Objectives The Supplier must provide specialist resources to maintain, scale and support the robust operation of AI tools and components, as well as the underlying data, AI and technical infrastructure used across AI teams. The Supplier will provide specialist resources to: • Scale AI solutions for department-wide use with reliability, security, and maintainability. • Ensure AI components and services are reliable, performant, and secure through strong data, infrastructure, and ML engineering practices. • Enable reuse of technical solutions across multiple use cases to simplify the overall technical landscape. • Recommend infrastructure approaches aligned with the Buyer’s technical guidance. • Ensure compliance with the Buyer’s engineering standards, governance frameworks, and internal assurance processes. • Monitor and manage performance, cost, and environmental impact of AI services, including LLM usage and infrastructure consumption. • Provide product and delivery management for live and mature services, including identifying user-driven improvements and managing enhancements. • Upskill DBT civil servants in operationalising AI, including core technical and infrastructure skills. • Provide guidance and set standards on how the Buyer’s DDaT teams can best deploy AI in their own services, • Continuously improve AI services and tools based on evidence-driven user insights. • Showcase AI Ops work to the wider DBT through demos and show-and-tells. Scope The supplier will be expected to deliver the following (which will be refined on a case-by-case basis as SOWs are agreed): Operationalisation and Scaling • Take AI Factory tools and components and operationalise them into robust, reusable, and scalable services on the Buyer’s infrastructure. • Maintain and enhance live AI services and tools, ensuring alignment with the Buyer’s standards as defined in agreed SOWs. Component Development and Management • Develop and iterate reusable AI components; maintain a shared component library on the Buyer’s infrastructure . • Partner with the Buyer’s digital teams to create components that integrate across multiple digital and data services, promoting efficiency and consistency. Infrastructure and Compliance • Implement Infrastructure as Code (IaC) and follow the Buyer’s engineering best practices as defined in agreed SOWs . • Engage with governance and assurance processes, including IRAP/TDA compliance. • Recommend infrastructure approaches aligned with DBT technical guidance. Monitoring and Sustainability • Provide service monitoring and enhancement to optimise performance, reduce waste, control costs, and minimise environmental impact. Documentation and Knowledge Transfer • Produce and maintain documentation for developers (both Supplier and Buyer) and the Buyer’s end users • Support the DBT AI and wider DBT DDaT teams in using AI-related infrastructure and components effectively. • Collaborate with the Buyer’s teams to ensure alignment with existing infrastructure and standards. • Deliver mentoring and pairing to build civil service capability in AI operationalisation. Continuous Improvement • Identify new features through user research, then design, develop, test, and evaluate enhancements to improve existing AI services. Please refer to the ITP (Appendix B) for more information on the required skills and roles needed for this lot.