AI-Enabled Legacy Modernisation Playbook

What this artefact is

A comprehensive analysis of the legacy application source code, extracting everything a developer needs to understand the system’s architecture, business logic, integrations, and data model.

How it’s produced

The application analysis is produced by the application-developer agent, which reads the entire application codebase — every source file, configuration file, and project file — to build a complete technical picture of the system.

What it contains

The application analysis has nine mandatory sections:

1. Application Overview

High-level summary covering:

• Purpose of the application
• Technology stack (language, framework, runtime)
• Solution structure (projects, layers, key folders)
• Dependencies (NuGet packages, third-party libraries)
• Configuration (connection strings, app settings, feature flags)

2. User Roles and Access Control

• Roles table listing each role and its permissions
• Authentication mechanism (e.g. Windows Authentication, forms-based)
• Authorisation approach (role-based, claims-based, custom)

3. Features and Capabilities

Named subsections per feature, describing what the feature does from a technical perspective and which source files implement it.

4. Workflows and Behaviours

Named subsections per workflow, documenting the sequence of operations the application performs for key business processes.

5. Business Rules and Validation

A numbered table (BR-xxx format) of business rules extracted from the code, each with:

• Description of the rule
• Where it is enforced (source file and method)
• Criticality (high, medium, low)

6. Domain Model

Entity subsections with:

• Property tables (name, type, constraints)
• Enumerations and their values
• Relationships between entities

7. Integration Points

Documentation of all external touchpoints:

• Web services and APIs consumed or exposed
• File I/O (imports, exports, file shares)
• Email notifications
• External systems and their connection details

8. Reports

Each report documented with:

• Type (e.g. Crystal Reports, SSRS, custom)
• Purpose
• Data sources
• Parameters
• Output format

9. Cross-Reference: Application to Database

Mapping between application concepts and database structures:

• Data access patterns (ORM, stored procedures, inline SQL)
• Entity-to-table mapping
• Stored procedure calls and where they are invoked from

Quality criteria

• All source files read and accounted for
• Business rules numbered and assigned criticality
• Domain model entities match the codebase
• Integration points documented with connection details
• Cross-reference section links application concepts to database structures

Example

The following snippet shows a technology stack summary for a fictional sample tracking system.

Technology Stack

• Language: VB.NET
• Framework: ASP.NET Web Forms (.NET Framework 4.0)
• Database: SQL Server with stored procedure-based data access
• Architecture: Three-tier with business logic library and web presentation layer
• Authentication: Windows Authentication via the agency’s Active Directory
• Key dependencies: Crystal Reports 13.0, Microsoft Enterprise Library 5.0

Business Rules (excerpt)