Reimbursement Intelligence

Context and challenge

Fraud and irregular reimbursement claims create massive financial risk for insurers. We mapped the current ecosystem (as-is) and designed an optimised future flow (to-be) to orchestrate automated fraud detection end to end.

• Current journey map
• Future journey map
• Success criteria agreed with the client

Mapping the ecosystem: metrics, variables and rules

I mapped the operational journey to spot bottlenecks, set success criteria and isolate the determining variables for the predictive models. This stage defined the system's priority rules and shaped the information architecture of the analytical dashboards.

• Solution A — Fraud and deviation score
• Solution B — Provider analysis
• Solution C — Image intelligence

Data prioritisation: effort × impact

Each candidate indicator was assessed with data engineers and scientists for development effort and business impact, producing a defensible analytical backlog — from ranking offending providers to detecting premature claims from newly accredited providers.

Strategic scope: user stories and data mapping

I translated the prioritised indicators into user stories connected to available data sources, making sure every screen came from a real business decision and from data that could actually be collected.

Low and high-fidelity prototypes

I built the claim analysis flow and the monitoring dashboards: member view, provider view, geographic ranking maps and cuts by period, service and portfolio.

• Reimbursement claim analysis flow
• Member intelligence dashboard
• Provider intelligence dashboard

Engineering handoff and evaluative research

Usability testing with the insurer's analysts guided dashboard refinement. A structured handoff with the engineering team kept the shipped product faithful to the design.