PortfolioFrom strategy through to production, applied to real problems
A selection of projects, tools, and applied work that reflects how I approach problems: structured thinking, systems design, and delivering working systems from strategy through to production code.
Projects
Frameworks and full-stack applications, each solving a specific structural problem.
myTrajectory
Job searching is fragmented. CVs, applications, interview prep, and career guidance all live in disconnected tools that don't talk to each other. myTrajectory connects them into a single workspace where every action informs the next, and AI works with the full picture rather than isolated snapshots.
Live AppBehaviour Intelligence from Web Analytics
Web analytics reveal what happened. This framework goes further, classifying visitors into behavioural states that directly map to actions, so teams stop debating dashboards and start making decisions.
Open SourceStructured Briefs for CRM Campaigns
An interactive, browser-based operating system for CRM teams. Connects high-level lifecycle strategy to day-to-day campaign execution.
Live AppArticles
Written perspectives on strategy, technology, and ways of working.
Balancing forcesThe Role of Product, and What Comes Next
A 2017 argument about why product exists to balance technology and commercial forces, originally presented at Racing Post and Tracr. Includes a postscript on how AI is dissolving the boundaries entirely.
Killing ideas systematicallyStart With the Problem
An open brief on agentic commerce: nine ideas generated, eight eliminated through structured invalidation. How reframing the question, understanding the domain, and testing assumptions systematically leaves you with the idea worth pursuing.
Speed vs coherenceVibe Coding Is Fast, but Implementation Sprawl Is Real
AI-assisted coding accelerates prototyping, but the same idea quietly appears in multiple places, each slightly different. This article proposes separating experimental code from production code, using prototypes to discover specifications through interaction rather than shipping directly from first drafts.
Signal vs volumeWhy Hiring Is Becoming Noisier
Modern hiring is structurally broken by volume. Candidates optimise and mass-apply to survive; employers filter until context disappears. The result is a system where visibility matters more than suitability, and quiet, neurodivergent, and older workers face disproportionate barriers.
Career Stories
Each story is built around a transferable principle: how the problem was diagnosed, what was proposed, what happened, and what it taught.
Medico-Legal Agency Engagement
A specialist consultancy was losing the majority of incoming enquiries. The problem was framed as needing better AI tools. The real issue was upstream: expert data that lacked structure, and a taxonomy gap between specialisms and incoming cases. This engagement reframed five pain points, prioritised one, and delivered a working proof of concept with two-stage AI matching and scenario-aware response generation.
Data before AIShawbrook Bank
The bank championed a product-led vision while its structure kept product managers focused within their own verticals. I proposed an incentive-driven shared services model built on voluntary adoption, developed through deliberate ground-up consultation before going to leadership.
Incentives vs structureAdvent International Labs: Unified Search
The Labs platform had three products with isolated search, making two of them effectively invisible to most users. I built the cross-linking logic within my own product first to prove the concept, then gained buy-in to roll out a unified search across the platform. Knowledge Center reached 50% adoption within weeks.
Proving then proposingAdvent International Labs: When a Mandate Is Not Enough
An attempt to introduce structured prioritisation at a team offsite broke down because the people whose work it would change hadn't been brought on side beforehand. The lesson was structural: a mandate gives permission to attempt buy-in, not buy-in itself. It changed how I approach change work: find the people who could cause it to fail, and engage them early.
Losing the room11:FS Foundry
When COVID hit, the anchor client stepped back and the rest of the product team was made redundant. I remained as the sole product manager and built a decision-tree roadmap to hold direction for the team through commercial uncertainty.
Mapping known unknownsTracr (De Beers Group)
The diamond traceability platform matched stones by carat weight, which many diamonds share. As the first product hire, I worked with the data science team to build a richer identifier combining 3D model matching with weight and other metadata, giving each stone a composite fingerprint reliable enough to make the pilot viable.
Finding the wrong modelExperian
The team had built a process for evaluating ideas before committing to them: drive fuzziness out of a proposition, test the cheapest question first, and delay investment until evidence justifies it. The process, and the thinking behind it, changed how I approach product development.
Testing before investingSitel / PlayStation
Agents shared knowledge through Outlook Express inboxes that regularly hit their storage limits. I built a contextual search tool in VBA: one input field that inferred what you needed from what you typed. It spread through organic adoption and led to a special projects role, an early lesson in constraint-driven building and letting usage make the case.
Building before permissionStudy Tools
Adaptive exam tools I built for my own study, designed around spaced repetition and mastery tracking, and shared for free.
Custom GPTTOGAF 10 Exam Coach
Adaptive multiple-choice practice for TOGAF Foundation and Practitioner. Tracks per-unit mastery, adjusts difficulty based on performance, and provides memory reinforcement for missed questions.
Custom GPTBCS Foundation Exam Coach
Adaptive multiple-choice practice for the BCS Foundation certificate. Same approach: per-unit tracking, progressive difficulty, and concise feedback grounded in official syllabus material.
Custom GPTBCS Practitioner Scenario Simulator
Generates realistic Practitioner-level business scenarios, lets you respond without interruption, then assesses your answer against BCS marking criteria with structured feedback on what to improve.