King: Fraud and Risk Tooling That Cut Investigation Waste

King's payment platform sat underneath millions of microtransactions across multiple markets, which meant fraud decisions had both revenue and support consequences.

The core issue was not missing data. We had signals spread across accounts, payment methods, registration details, and support history, but investigation work was fragmented and slower than it needed to be. Fraud handling consumed support time, internal review time, and still left room for inconsistent judgement.

I worked with the team on internal risk tooling aimed at making investigations faster and more consistent.

The focus was:

• Bring the most useful fraud signals into a more practical review flow
• Reduce manual comparison work across duplicated or suspicious accounts
• Improve decision support for the people handling risk and support cases
• Target operational leverage, not just more reporting

The result was material:

• Fraud reduced by roughly 40%
• Around 50 support hours per month saved
• Investigations became faster and easier to reason about
• Internal teams had a clearer basis for action instead of piecing evidence together manually

One useful example was a QlikView-based investigation view I put together when we had the data but not the delivery capacity for a full backoffice build. It allowed practical filtering and comparison across signals like username, email, IP address, registration date, and number of payment cards. It was not the final state of the system, but it gave the team a working tool at the moment it was needed.