Methodology
The approach
How I work, what I deliver, and what you will never get from an engagement with me.
Operating principles
Four rules I don't break
I don't recommend tools I haven't run in production
Every model, every framework, every deployment pattern I suggest has been tested on my own infrastructure first. I know the failure modes because I've hit them.
I document everything so your team owns it after I leave
The goal is independence, not dependency. Every system comes with architecture documentation, runbooks, and a trained internal team.
I work in weeks, not quarters
Discovery in days, not months. First working system within weeks. If something isn't delivering value fast, we change direction or stop.
One consultant. No handoffs to junior staff
You talk to me, I do the work. No account managers, no junior consultants finishing what I started, no telephone game between the person who sold the project and the person building it.
The process in detail
From first call to production
Every engagement follows this structure, adapted to scope. Timelines are typical — some projects are faster, some need more discovery.
Week 1–2
AI Audit
Map the landscape, identify the starting point.
I interview stakeholders across business, IT, and compliance. I review your existing tools, data architecture, and security posture. The output is a prioritised map of 3–5 AI opportunities ranked by impact and feasibility — not a 60-page slide deck, but a working document your leadership team can act on.
Deliverable: Prioritised opportunity map with impact/feasibility scoring
Week 2–3
Governance & Privacy
Align AI with your compliance reality.
Data classification, access controls, processing agreements, EU AI Act risk assessment. I work with your legal and IT security teams to build a governance framework they can sign off on — because the fastest way to kill an AI project is to skip this step and get blocked three months later.
Deliverable: Governance framework approved by legal and IT security
Week 3–8
Pilot Implementation
Build the first working system.
Production code on your infrastructure, integrated with your existing systems. Not a demo, not a prototype — a working system that handles real data and delivers real output. We validate assumptions with actual users and iterate weekly.
Deliverable: Working system in production with real users
Week 8+
Scale & Training
Hand it over completely.
Documentation, training sessions, operational runbooks. Your team learns to operate, monitor, and evolve what we built together. I make myself unnecessary — that's the point.
Deliverable: Full documentation and a trained internal team
Full transparency
What you won't get
Slide decks that gather dust in SharePoint
Vendor recommendations driven by kickbacks or partnerships
Solutions that require my ongoing involvement to function
Broad AI strategies with no specific implementation plan
Recommendations I haven't validated on my own infrastructure
Ready to start?
A 30-minute conversation, no obligation. We map your situation and you leave with clarity on the next step.
Book a Sparring Session