Nine years in product, and the operations desks before it

Clearon handles digital coupons, gift cards and payment clearing for around 6,000 stores. One development team of six developers and two QA, several legacy systems, and a backlog driven from above.

Three initiatives in five and a half months, definition to delivery. Campaign sendouts with the consent handling that goes with them, a prepaid gift card subscription flow with balance checks against Visma, and a fulfilment model for digital value objects, built for Triss lottery delivery but made to carry other products later. Alongside that, early discovery on turning the main products into a SaaS offering.

There was no spare UX or engineering capacity, so I prototyped the answers myself with AI assisted tools rather than wait for a slot. A dashboard package built from the backend API contract, an internal calculator for Customer Success, and funnel and consent tracking around sendouts.

A loan broker, where conversion and compliance have to coexist. My focus was business loans and insurance. Long journeys, complex integrations, and an expensive failure mode when a flow breaks.

I owned the insurance flows day to day with a product designer, effectively a pair working straight with development. The accident insurance was the awkward one, since it had to win attention against Lendo's own existing insurance in the same flow, so we ran concepts as live A/B tests rather than picking a design and defending it. We were still tuning it when the team was disbanded. No numbers for that stretch, because I was not tracking any, and I would rather say so than estimate one now.

The Monto.ai integration is the one worth telling. It worked, and then it came out again. Usage stayed low because nobody had established that a business owner applying for a loan often cannot log into their own accounting software. That access sits with the accounting firm. The build was fine. The question nobody asked before it was not, and I have asked it on every integration since.

Also a Focus API integration and autogiro work, and a move from Scrum to Kanban once the work shifted to long running integrations and sprints were costing more than they returned.

Payments were never a career move. They arrived attached to the risk and fraud work and I learned them on the job. I had the technical half through the royalgames.com years, a colleague had the commercial half, and when she left in 2018 I took hers too and moved back to Sweden from Malta.

The operations title went to someone new, but much of the work stayed with me. Little of what I knew about that platform had ever been written down. By then the two of us were the business, with support outsourced and a single developer left at the end.

Tens of thousands of microtransactions a day on king.com, thousands on royalgames.com. The US and Germany were the largest markets, and which methods and providers were available differed market by market on one platform. PayPal, Klarna, Paysafecard, Neteller, Payvision, and the fees, refunds and chargebacks that come with them.

My fraud titles had ended years earlier. I kept fraud and risk anyway, because dropping it would have cost more than keeping it.

I joined King in 2008 and spent the first years on the payment support, risk and fraud desks, including a spell leading a small second line team. By 2015 I owned product and operations for royalgames.com, because I had worked with the platform since 2008 and knew it as well as anyone.

It was a real business and a shrinking one. The 2015 accounts, the year I took it on, show around 12 million euro in revenue, down from close to 16 million, with the decline expected to continue as players moved to mobile. I was the only product person on it, working with three developers, later two. Nobody was funding a turnaround, so the job was deciding what still deserved effort.

The fraud signals existed but nobody could look at them together, and there was no capacity to build anything into the admin system. I designed a QlikView investigation solution and specified its search logic. A QlikView specialist implemented it, since I had no build access. An estimated 40% less player-account fraud, around 50 support hours a month freed, at roughly a tenth of the projected cost of building it properly. There is a longer write-up.

Two smaller ones. First line support was deciding case by case what needed fraud review, so I mapped the logic into a branching flow, drew it in Word and shared it as a clickable PDF. It cost nothing and cut processing time, errors and incorrect escalations by an estimated 45%. And approved withdrawals still took three repeated actions after the checks, so I sent them to a batch list instead. Across 300 to 500 a day that removed roughly 600 to 1,000 clicks, with the final payment control untouched.