Delivering Validated Training Courseware for UK Royal Engineers at Machine Speed

12 May 2027
New Horizons Keynote Theatre
Capability at pace

Developing training courseware for the Royal Engineers is complex and resource-intensive, requiring the integration of technical engineering standards, Defence policy, military doctrine, equipment documentation, and training requirements. Subject Matter Experts (SMEs) must analyse extensive and sometimes overlapping source material before converting it into effective, accurate and auditable training. In partnership with MKC Training, Mission Decisions is investigating how Large Language Models (LLMs) can accelerate this process while preserving human authority, source traceability and doctrinal integrity.
Addressing Challenges
Unlike civilian training environments, Defence courseware must distinguish between industry and engineering standards and the principles, policies and practices established through Defence doctrine. Manual analysis of these sources can be time-consuming and difficult to scale, particularly when documents contain ambiguity, duplication or conflicting requirements. The research addresses how AI can support the analysis and structuring of this information without automatically resolving discrepancies or becoming the authority of record.
Methodology
Mission Decisions' Analytical Engine orchestrates multiple, specialised LLMs within an automated courseware-development pipeline. Inputs include source documentation, applicable standards and Defence doctrine, learner profile, existing knowledge, learning objectives, learning points and course duration. The Engine generates a structured syllabus which subsequently drives courseware production through a consistent prompt architecture. Multi-level document processing controls the context provided to the LLM, while a secondary verification model cross-checks generated content against identified authoritative sources, highlights ambiguities and contradictions, and identifies areas requiring SME review. SMEs remain fully human-in-the-loop, retaining responsibility for interpretation, validation, approval and amendment.
Results
The research demonstrates the potential to significantly reduce the manual effort involved in transforming complex technical and doctrinal material into structured training. The framework provides a repeatable approach to syllabus development and content generation while maintaining source references and highlighting areas where expert judgement is required. Where discrepancies occur between technical standards, source documentation and Defence doctrine, these are flagged for review rather than automatically modified, preserving traceability and document integrity.
Implications
The approach provides a potential pathway towards scalable, AI-assisted Defence training development. By combining automated analysis and generation with layered verification and human assurance, the framework can increase SME productivity while supporting consistency, traceability and controlled adoption of AI within Defence training. The underlying principle is that AI analyses, structures and challenges; the SME validates, interprets and authorises.

Chairperson
Carlos Montans
Carlos Montans, Head of Training Centres Services - Airbus Defence and Space
Speakers
Taja Hillier CStat
Taja Hillier CStat, Chief Data and AI Officer - Mission Decisions