top of page

AI Systems

AI doesn't work when it is bolted on. AI works when it is introduced surgically, with clear understanding of the overall system. That is what I do. I come from fifteen years in journalism and I bring that discipline to every system I build: learn the who what where why, then build.
 

I design and build AI content infrastructure that works. My deepest experience is in content systems for learning organisations and for podcasts and media projects.

I worked with:

Case Study: reducing prep time by 75%

A global professional education company was scaling its enterprise business — selling customised learning programmes to large organisations. The problem was in the sales process: every client proposal required a specialist to manually assemble course content, match it to the client's needs, and produce a formatted deliverable. It was slow, inconsistent, and impossible to scale.


 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I designed and built a content system from scratch. At its core: a GitHub-backed repository with a clean architecture per course, canonical metadata files that the system could read and reason about, and a Claude API integration that transformed a client brief into a formatted, personalised proposal in real time. 


The result was a reduction of over 75% in manual preparation time, validated by the enterprise team lead on the day of handoff. 

Consulting Methodology

Before any system gets built, I map the knowledge: how things are currently done and where things can be more efficient. This is why when I leave, the results stay — they are not 'added on', they are embedded in the actual work processes. Architecture, logic, editorial rigour

bottom of page