Defra AI Legacy Modernisation Playbook

Version 1.0.0

Purpose of this Playbook

This playbook provides guidance on how to modernise legacy systems using AI technologies. It is designed for internal use within Defra and is updated regularly with new practices and lessons learnt.

Note: All examples provided in this playbook are illustrative and intended to show contextual possibilities only.
They do not represent confirmation of existing systems or projects currently in use by Defra. Any examples of potential AI tools and framework use (e.g. LangChain, Azure AI Studio, Semantic Kernel, OpenAI Assistants) are illustrative only. Always consult the Defra AI Tools Authority or your technical governance lead before selecting or implementing any AI technology to ensure alignment with internal standards, approved platforms, and procurement policy.

Table of Contents

1. Welcome
2. Responsible Practices
   • Ethics
   • Sustainability
3. Getting Started
   • Four Pillars
   • Workflow
4. Discovery and Analysis
   • Documentation Gathering
   • Requirements Building
5. Pathways
   • MS Access
     • SSMA Migration
     • SQL Refactoring
     • Document UI
   • SQL Server
     • Schema Extraction
     • Changelog Refactoring
     • MongoDB Decision Point
     • MongoDB Migration
     • OpenAPI Specification
6. Lessons Learnt
7. Appendix
   • Defra Approved Tools
   • Prompt Library

Overview

Defra manages environmental protection, food production and rural development across the UK. The department works through multiple sub-departments and arms-length bodies all of which have legacy systems that can benefit from modernisation.

Next: Responsible Practices →