I spot where the work could run better, organise a way through, and create the solution. I build with AI, from scoping the problem to directing the build.
I work in facilities & events, and I spend the rest of my hours building AI tools. That started as automating my own admin. It turned into a tool I proposed at work and built with the teams who would use it, which has now run a live trial and been endorsed by the branch that ran it.
I do not write code unaided. I scope the problem, direct AI to build it, then read, run and validate what comes back — and I can defend every design decision in the tools below. Coming from operations rather than engineering is the useful part: I pick problems that cost the business something, because I have done those jobs.
Work project · live trial
A sales team picked who to call next from spreadsheets, memory and experience. This proposed tool scores every potential customer and hands each person a ranked list, with the reason for each name attached.
Personal project · in daily use
An extension of my eyes, hands and memory. It has grown a lot over time and now has dozens of tools behind it that help me run my life, my research and my development. It knows what I am working towards, keeps my week in front of me, and handles the parts I would otherwise put off.
How it was built — Client targeting tool
Deciding which clients to approach next used to rely on knowing the region, going through several data sources by hand, and years of experience. Two people with the same list would pick different targets for different reasons. I worked out that we could build our own data source and analyse it for a competitive advantage, then built it with the teams who would use it: it scores every account against external data and hands each team a ranked list with the reasoning attached.
During the first trial, 81% of new booking days came from accounts it flagged — 3.6× the comparison set. Endorsed by branch management, and the teams now use it day to day.
It is now in its second-stage trial, so keep your eyes peeled for updates.
How it came about — My own AI assistant
It started as one script that updated a spreadsheet of my savings, because I kept forgetting to. Everything since has been the same move repeated: notice something I do by hand every week, hand it over, and keep it only if it earns its place. What is left is dozens of tools and one assistant that knows my goals, my constraints and how I like to be argued with.
It is plugged into the things my week actually runs on. It reads and writes my calendar, so it plans around what is already there rather than asking me. It pulls live prices from my broker, my exchange and market feeds. It keeps my notes and research in one place it can search exactly, rather than guessing at what looks related. And it writes what it needs to write — a designed document, an image like the one on the card beside this, research, insights and, most importantly, ice cream recipes.
The honest limit: it is built for exactly one person, which is what lets it be this specific and also why it would not work for you as it stands. The failure I actually design against is silence — a dead feed leaving yesterday’s numbers looking like today’s — so how fresh a number is gets checked before it is ever shown to me.
Oct 2022 — present
TeacherActive
Company-wide health and safety, supplier contracts, and office relocations from teams of 4 up to sites of 30. Negotiation and process change here produced £24,230 in documented savings across 2024–25, before any AI was involved. It is also where I proposed the client targeting tool and built it with the teams who now work from it.
Oct 2021 — Oct 2022
Wagamama
Service delivery under real time pressure. Employee of the Month three times.
Apr — Sep 2021
Kovacs Group
Warehouse operations and inventory accuracy, measured by the hour.
Oct 2018 — Dec 2020
Java Roastery
Daily operations, inventory control and staff training. My first job running a site rather than working in one.
Jul — Oct 2018
CastleYard Tearooms
Event delivery, start to finish. Where the events half of what I do now started.