AI is changing how software is built. But adding AI to a product because it’s popular rarely creates value. The real question is where AI genuinely helps — and where careful engineering matters more.

I help businesses put AI to work inside the products and workflows they already use: assistants trained on your own content, document analysis, and automation that takes repetitive work off your team’s plate. The goal is always practical — reduce manual effort, improve decisions, and keep the results reliable.

Where AI adds real value

  • Knowledge assistants trained on your documentation and policies
  • Customer support that answers common questions and escalates the rest
  • Document analysis — extracting structured data from contracts, invoices and reports
  • Intelligent search using natural language
  • Workflow automation for repetitive administrative tasks

Engineering, not demos

Anyone can build a demo with AI. Production systems need more: context management, retrieval, data privacy, model selection, caching, monitoring, evaluation and sensible fallback behaviour. I build AI features that hold up when real users arrive — and I’m honest about where a traditional solution would be more reliable.

My experience

I’ve spent the last couple of years using modern language models daily in real development work, and building with approaches like retrieval-augmented generation, function calling and the Model Context Protocol. That hands-on use is what lets me tell you where AI helps and where it gets in the way.

[NEEDS: a shipped client AI feature to reference here, if available]

Thinking about adding AI to your product? Get in touch.