Services · how I can help
Put AI to work.
Then find out if it worked.
I help small organisations and teams with three questions. Each starts with one small, fixed first step, and each ends with a written read of what the evidence can and can’t support.
01Lead offer
AI that people use
- Who it’s for
- Small firms, practices and teams of 5 to 50.
- What we do
- A discovery session: walk through a real working day, find the handoffs, rank the options by impact, effort and risk, and agree one first step. A half-day team workshop can follow.
- What you leave with
- A ranked shortlist of where AI fits your team's real day, and one first step to try next week.
- First step
- 90-minute discovery session · Price on enquiry
02Available now
Did it work?
- Who it’s for
- Founders pitching funders, not-for-profits, program managers.
- What we do
- Evidence plans, KPI and survey audits, pilot design and a written readout of what the results can and cannot support.
- What you leave with
- A two-page read: who needs to believe what, what each would accept as proof, and what to collect first.
- First step
- Evidence triage call and a two-page read · Price on enquiry
03Available now
Is the AI any good?
- Who it’s for
- Teams building or buying AI agents, and people who make benchmarks.
- What we do
- Evaluation audits, custom benchmark design and checks on whether an AI judge can be trusted for your task.
- What you leave with
- A written audit: where the evaluation holds up, where it does not, and the test cases that would settle it.
- First step
- Evaluation audit · Price on enquiry
See it coming. A weekly signal feed built around one client's decisions. Pilots by arrangement. Ask about a pilot ↗
What happens when you get in touch
Small steps, and you decide each one.
- A 20-minute call. You describe the problem in your own words. I say whether I can help, and if I can't, who might.
- One fixed first step. Scoped and priced before it starts, so you know what you are buying.
- A written read. What the evidence supports, where it stops, and what I would look at next.
- Your decision. You choose what follows, if anything. No retainer is assumed.
What every engagement shares
Every engagement ends with a written read of what the evidence can and can't support.
How I work
Four steps, every time.
- Name the claim. Say exactly what you want to be true, and who needs to believe it.
- Read the signs. Find what can actually be observed, and what each sign could mean.
- Grade the warrant. Judge how far the evidence carries the claim, and where it stops.
- Design the next test. Choose the cheapest observation that would change the decision.
Who you would be working with
One person, and what I will and won’t claim.
My background is in psychological science and organisational psychology, with graduate training in measurement. I build with AI every day, and I measure what it does.
More about me and the name ↗- I am not a registered psychologist, and none of this is clinical or psychological assessment.
- I use AI throughout the work. I remain responsible for the questions, the claims and the conclusions.
- Client names and material stay private unless the client agrees to be named.
- I will tell you when the evidence does not support what you hoped it would.
Not sure which of these is yours?
Most people aren’t. Tell me what is happening in your own words and I will say where I would start.
Book a 20-minute call ↗