Problem

Analysts rebuilt one person from 8+ systems, and risk scores came with no reason.

What I did

I owned research, IA and UI, and turned risk into a map that traces every rating.

Impact

One search, one profile and one shared way to read risk.

8 → 1

systems to one search

X → Y min

per case review, before and after

3

risk levels, one rule set

Overall risk map showing which records feed the risk level
02 · PROBLEM

A score nobody could explain

Opening a case meant switching tools and copying details by hand. When a rating said high, the analyst still had to dig through every system to justify it.

“Show me why, not just how much.”

Analyst, discovery interview

03 · IDEAS I REJECTED

Three sketches, two dropped

NameSourceStatus

✕ One big table

Complete, but no story. Rejected in the first review.

JobsAssetsTravelFamily

◐ Cards per source

Clearer, but risk still hidden. Kept as the overview.

HIGHDebtTxnCourtShipFin.VisaFamily

✓ Risk as a map

Every rating shows its reason. Chosen.

04 · KEY FLOWS

Problem, then solution, for each flow

FLOW 01 · PROFILE OVERVIEW

Problem

Tables buried the story of a person.

Solution

One card per source with a status badge and a link into the record.

Profile overview with summary cards for each record source
Profile screen with a connected risk map around the subject
FLOW 02 · RISK GRAPH

Problem

Scores had no reason behind them.

Solution

A map inside the profile, so analysts never leave the person.

05 · IMPACT

What the data intelligence platform changed

X%

less time per case review

8+

record sources in one profile

1

dark design system for every module

06 · WHAT I LEARNED

Show the why.

Trust came from traceable reasons, not a better score.

Test earlier.

I would put the map in front of analysts in week one, with their real cases.

There is more to this work than fits on a page.

MORE WORK
AI FORECASTING · 5 MIN READ

From bookings to officers on duty →

ALL CASE STUDIES

See more of my work →