AI Product Architect
I design the systems and interfaces that turn ambiguous AI ambition into things people actually understand, trust and ship.
A responsible AI flow continues through execution, verification, recovery, contestability and learning. Approval is a transition, not the finish line.
Featured Case Study
Solve Axis is a concept for an AI-native enterprise supply-chain modelling product — organised around how analysts actually reason about a solve, with a designed accountability layer that makes AI safe to trust on a high-stakes decision.
Narrated walkthrough — best watched with sound on.
Work
SS-1 Signal Shell is a market intelligence engine for product teams. It monitors competitor changes, filters noise, generates structured strategy briefs and asks a human to approve any update to the product's strategic record. The project explores how AI can support competitive judgement without turning every external signal into a reactive backlog item.
Swarm Lite is an AI product strategy prototype that helps teams test strategic questions before turning them into roadmap work. It generates representative market personas, runs structured huddle sessions, produces decision-ready strategy briefs and routes consequential actions through human approval. The project explores how AI can challenge assumptions earlier without pretending to replace real research.
TX-1 Terminal Explorer is a prototype for AI that earns the right to act. It diagnoses supply chain optimisation failures, validates proposed fixes in a dry run, and asks a human to approve the change before anything is committed. The project explores what trustworthy agentic AI should look like in serious enterprise workflows: visible reasoning, bounded autonomy, human ownership and a full audit trail.
A useful AI product does more than produce an answer. It helps evidence, judgement, authority, action and accountability move safely through one decision.
A human approval gate can work perfectly while the surrounding AI system remains unaccountable. The real design challenge is the chain of capabilities around the decision.
Most human-in-the-loop AI systems produce logs, not decision records. A log tells you a button was clicked. A real record tells you what the human knew, decided and what happened.
AI agents can make technically valid recommendations that are still wrong because they lack business context. More data does not solve it. Better context does.
