Approach · AI-assisted modernisation

Make legacy change easier to understand, deliver and verify.

I help engineering teams structure modernisation work around clear requirements, bounded agent tasks and human review. The starting point is the behaviour the business needs to preserve.

Business contextClarify what must stay true as the system changes.
Structured deliveryGive agents clear context, scope and instructions.
Reviewable changesAgree how each increment will be checked.

Start with the behaviour that matters

A modernisation plan needs to make business rules, dependencies and open questions explicit. My approach starts with defining the intended behaviour and agreeing acceptance criteria before expanding implementation.

A workflow for controlled progress

I use a specification-led approach to connect requirements, implementation tasks and verification. The scope and checks should fit the system and the team.

  1. UnderstandIdentify behaviour, constraints and open questions.
  2. SpecifyAgree requirements and acceptance criteria.
  3. PlanDefine bounded tasks and dependencies.
  4. ImplementWork in increments that can be reviewed.
  5. VerifyCheck results against the agreed criteria.

Give agents a clear working environment

I help teams define the context, reusable instructions, tool access and task boundaries for AI-assisted work. The workflow should make it clear when an agent can proceed and when a person needs to resolve a question or approve a decision.

Define verification before scaling up

For a modernisation engagement, I would agree checks such as:

The aim is to give the team evidence for deciding whether an increment is ready and what needs attention next.

This page describes a consulting approach, rather than a specific client engagement or its results.