Senior Associate
Bringing AI into high-stakes commercial litigation - as a tool, not a substitute.
I am a Solicitor Advocate at Stewarts specialising in high-stakes cross-border litigation, regulatory investigations and complex fraud. My practice is unusually dual-skilled: alongside the law, I work in Python, generative AI, blockchain and Linux. I hold a Certified Junior Penetration Tester qualification - an ethical-hacker grounding that translates surprisingly well to how modern commercial disputes need to be approached. I spearheaded one of the first-ever GenAI solutions in the Commercial Court. The work included building a custom OCR and GenAI processing pipeline over 100,000+ PDFs and an in-house machine translation tool built on top of it. Earlier, I was part of the team that delivered the first fully virtual trial in the Commercial Court - a hearing widely recognised by judges, experts and the press as a watershed for remote justice. I now sit on Stewarts' GenAI Committee, guide the firm's AI strategy and have independently developed three AI software tools deployed inside the firm.
Most of the legal-AI conversation starts and ends with the LLM. That framing misses the point. The biggest gains I have seen in commercial litigation have not just come from asking a model to "do the legal work" - they have come from building proper engineered pipelines where AI is one component among many: deterministic OCR, classical processing, structured data extraction and a generative model placed precisely where its strengths are an asset rather than a liability. That orientation comes partly from the ethical-hacking side of my training: you build for the actual problem, not for the fashionable answer to it. And in the context of commercial litigation, where evidence has to be defensible and outputs auditable, that distinction matters. The most useful question is rarely "can AI do this?" — it's "what's the most reliable, reviewable way to solve this and where does AI earn its place?"