I build the systems. Then your team runs them.

I am Conrad. I build production AI systems for businesses, on top of the software they already run. Right now those systems are drafting purchasing for approval, reconciling sale-or-return stock for the UK's leading leisure centres, and landing a prospecting briefing on a sales team's desks every Monday.

I am a problem solver first. Give me a process or a problem and the sharpest thing I do is the diagnosis: I map exactly where AI streamlines, assists or automates it, and where it should not go. I like the problems with no tidy off-the-shelf answer, the ones you solve by building from scratch or coming at it sideways. That map is day one of every build, and it is why the systems I ship get used: they start from how the work actually happens, not from what the technology can do.

Before any of this, I founded and ran a business. I started CrewPass to close a gap I had lived inside the superyacht industry, and running it taught me what actually moves a company: cash, time, risk, and the drag of admin that quietly eats people's days. I watched that manual work pull time away from the people doing it, then saw what changed when you put the right tools in their hands and streamlined the operation around them. That is where the passion came from, and it is what I do for other businesses now.

Somewhere along the way I worked out which part I actually enjoy: sitting with the people who do the work, finding the process that eats their week, and fixing it with them. Not writing a report about it. Some weeks that means wiring a 20-year-old back-office system into a clean dashboard. Some weeks it means building a custom engine from scratch for a problem nobody has packaged software for. And every build ends with your team shown how to run it, because a system only earns its keep when the people who own the work own the system.

I have built across systems of every generation, legacy and new, from twenty-year-old ERPs to greenfield stacks, so almost nothing about SME software surprises me any more. Under the hood I use the full modern AI toolkit: custom MCP connectors, retrieval-augmented generation, indexing and embeddings, automated grading and evals, agentic memory, and self-improving agents that sharpen the longer they run. I reach for whichever the problem actually needs. I am based in the UK and work with businesses anywhere: remote-first, on site when it matters.

Away from client work I mentor at Bournemouth University, and I am happy to speak to local business groups about what AI genuinely does for an SME.