Product Manager · London
I turn hard, ambiguous problems into products people use.
I’m a product lead based in London with a decade building across marketplaces, e-commerce, and operations. Taking complex opportunities and shaping them into beautiful, shipped products that moves the dial and puts a smile on customers' faces. Most recently leading product at Motorway. I’m currently consulting independently.
Experience
Productising a leadership-coaching function for a company serving Monzo, Deliveroo, and Unilever. Turning a bespoke, consultant-led offer into a packaged, repeatable product with AI-assisted tooling.
Led the Automation Squad shipping model-backed verification: computer vision for document validity, AI-assisted identity cross-referencing. Reduced the third-party ops team by 30%. Built a lead-prioritisation engine that lifted conversion by 20%. Managed 2 PMs.
Architected CRM integrations with Zendesk and Freshsales, enabling real-time data sync across the customer lifecycle. Established the measurement system for automation features that remains the squad’s source of truth today.
Joined pre-scale. Built the self-service post-sale journey combining triggered communications with a personalised portal, enabling 70% of customers to accept their sale without agent intervention.
Built a just-in-time inventory system reducing warehouse costs by 25%. Introduced an A/B testing programme that lifted add-to-basket from listings pages by 10%. Defined the architecture for a modular, event-driven platform.
Site Merchandising Manager and Content Editor at Swoon Editions. UK Publishing Manager at nCrowd.
Case study · Motorway
Automating identity and document checks.
Lead PM, Automation Squad · Feb 2023 – Mar 2026
A small squad replaced manual document and identity checks with model-backed verification, reducing the third-party ops team by 30% without weakening fraud detection.
01 — Context
Every sale on Motorway depends on two checks: that the seller is who they claim to be, and that the vehicle documents are genuine. For years a third-party operations team reviewed documents and cross-referenced identity by hand. This was accurate but slow and expensive, and it scaled linearly: every extra sale meant another minute of human review.
02 — The problem
As listing volume grew, manual verification became the bottleneck. Cost rose with volume, review times stretched and caused sellers to drop off, and accuracy varied between reviewers. Hiring more reviewers would have kept those costs scaling indefinitely, so we set out to identify which checks needed a person and which did not.
03 — Constraints
- • Fraud detection could not get weaker.
- • Compliance and ops had to understand and trust the system, so a black box was not an option.
- • Decisions had to be auditable after the fact.
- • A small squad: 2 PMs, engineering, and data science.
04 — The bet
We treated verification as a triage problem. Computer vision validates document authenticity and AI-assisted cross-referencing confirms identity, each producing a confidence score. Only edge cases are routed to a human reviewer, so reviewers spend their time on the cases that need judgement.
05 — What we shipped
- 01 A CV document-validity model catching tampered and mismatched paperwork.
- 02 AI-assisted identity cross-referencing across data sources.
- 03 A reviewer console showing the confidence score and the reasons behind it.
- 04 A measurement system that is still the squad’s source of truth today.
06 — Outcome
What I’d do differently
I’d build the reviewer console earlier. We first treated it as a thin layer over the models, but it turned out to be the most important part of the product. Building the models was relatively straightforward; getting ops to trust them enough to act on their output took longer, and the console was where that trust was built.
Side projects
Things built for the love of building.
Projects I build in my own time, usually somewhere between craft and hardware.
Contact
Get in touch.
I’m a London-based product lead, currently consulting. I’m interested in operationally heavy and ML-shaped problems, with teams that want to ship.
What I’m open to
- • Senior / lead PM roles at scale-ups solving hard operational or automation problems.
- • Productising bespoke, consultant-led services into repeatable products.
- • Selected product and AI-tooling advisory.