Approach

The five questions I work through before anyone picks a technology. The same questions drive consulting and implementation work.

Why the questions come first

Most data and AI programmes stall not because of the technology, but because they skip the questions and jump straight to tools.

After 25 years designing and building data platforms across Oceania and Europe, I work through five questions in a fixed order. It keeps the business goal and the technical reality attached to each other.

The 5 Questions Framework

Every data and AI programme must answer five questions, in the right order:

1. WHY → What business outcomes do you need?

What decisions or capabilities require data or AI? What business value are we creating? This defines priorities and success metrics before any technical discussion begins.

2. WHO → What team capability do you have?

What skills does your organisation have today? What structure fits your business model? Platform and AI choices depend on team reality, not vendor features.

3. WHAT → What data do you actually need?

Given your objectives and team capabilities, what data is feasible to capture and use? Where does it come from? What quality thresholds matter? Most AI failures are data failures, so this step decides what any model can deliver.

4. HOW → How do you deliver it?

What platforms, architecture, and processes fit your answers to WHY, WHO, and WHAT? Which models, what guardrails, and how do you evaluate that the AI actually works? Technology choices come fourth, not first. Governance enables rather than restricts.

5. WHEN → What's the roadmap?

What sequence makes sense given dependencies? What are quick wins versus strategic foundations? How do you phase investment, and how does a pilot earn its way into production?

Why This Order Matters

Most consultants start with HOW (platform selection, architecture patterns) without understanding WHY (business objectives) or WHO (team capabilities).

The result? Snowflake for a SQL-light team. Databricks for simple reporting needs. RAG pilots on data nobody can trust. Millions invested in technology that doesn't fit.

My approach starts with outcomes, assesses organisational readiness, validates data feasibility, then recommends technology that actually fits your reality.

How we work together

The same five questions drive both kinds of engagement.

Consulting

I work the questions with you and hand back a decision, an architecture, and a plan you can execute.

Implementation

I work the same questions, and then I build it: embedded in your team or leading the build.

Work with me

Email me, or connect on LinkedIn.