I start small and scale only what works. We agree what success looks like before any results come in.
The steps
Three steps, with a stopping point after each
Step 1
AI Opportunity Map
A short review of where your team's time goes and which jobs AI could take on. You get a ranked list with the value of each job and a plain account of where your data would go.
What you get
Interviews with the people who do the work
A ranked list of jobs with the value of each
A data map showing what would go where
About two weeks · fixed fee, quoted per engagement
Step 2
Gated pilot
One question and one pass mark, agreed in writing before any results exist. If the answer is no, you pay a reduced final balance and keep the findings.
What you get
A written pass mark signed before work starts
A working test on your real data
A clear yes or no, with the evidence
Three to four weeks · fixed fee, quoted per engagement
Step 3
Operator retainer
Once a pilot works, I build AI into one department at a time. A written permission ladder sets what the AI may do on its own and what stays with your people.
What you get
One department at a time
A permission ladder you sign off
Monthly review of what the AI did
Monthly, after a three-month minimum term · fixed fee, quoted per engagement
Most clients start with the map. You can come to me with a pilot already in mind.
How I work
AI earns access one rung at a time
Each piece of work has a written brief, an explicit out-of-scope list and a finish line you can check.
AI starts with read-only jobs such as a morning brief or a weekly summary. It earns the right to change live systems step by step, and some decisions always stay with your people.
A separate review checks the work against your real data before anything ships.
Rung 1
Reads and reports
A morning brief or a weekly summary. The AI can look but changes nothing.
Rung 2
Drafts for approval
The AI prepares the email or the update. A person checks it and sends it.
Rung 3
Acts within limits
Routine changes the AI may make on its own, inside limits written down in advance.
Always yours
Stays with your people
Decisions you choose to keep with a person, whatever the AI could do.
My data promise
Your data stays yours
Light passes through. Your data doesn't.
I choose the model that fits the data. Sensitive work runs on private open-source models on your own infrastructure, and the rest can use enterprise cloud services that don't train on your data. Often one solution combines several.
You own what gets built, and your data never becomes someone else's training material.
Is it a fit?
Who this works best for
A good fit
Owner-led businesses where people's time is the constraint
Teams with data too sensitive for public AI tools
Owners who want a clear yes or no before spending more
Probably not a fit
Companies looking for a chatbot on their website
Projects that need a large in-house engineering team
Work where nobody can say what a good result looks like
Clients
Who I've worked with
Clients so far include a music-tech start-up in the UK and an educational toys company.
Questions
Before you book
Which AI models do you use?
Whichever fits the job and the data. I am provider-agnostic, and one solution often combines a private open-source model with a cloud service.
Who owns what you build?
You do, along with your data.
What happens if the pilot fails?
The pass mark is written down before any results exist. If the pilot misses it, you pay a reduced final balance and keep the findings, so a clear no is still a finished piece of work.
Do we need technical staff?
No. I work with the person who owns the process day to day, and I handle the build.
How are prices set?
Each step has a fixed fee that I quote per engagement before it starts.
Do you work outside Australia?
Yes. Clients are in the UK and Australia, and I invoice in GBP or AUD.
Start with a map
A 30 minute call is enough to tell whether a map is worth doing.