Research
30+ analyst interviews
Scope
90+ alert types · L1 & L2 analysts
Outcome
Next-gen platform · GenAI integration

The problem

Every day, trade surveillance analysts sit down to do critical work — monitoring markets, reviewing communications, identifying potential misconduct across more than ninety alert types. The tools they use to do it are failing them.

Legacy systems at end of contract. Fragmented workflows. Missing data. And the defining frustration of analyst life: swivel chairing — the constant, exhausting act of jumping between four to five different systems just to disposition a single alert.

Research scope
Analysts interviewed
30+
L1 and L2 levels
Alert types covered
90+
Across surveillance scope
Systems per alert
4–5
Current state · L2 analysts
Analyst levels
L1 / L2
Both researched separately
The core finding

The best, most experienced analysts were spending the majority of their cognitive energy on process overhead rather than actual surveillance judgment. The system was working against the people it was supposed to support.

"Four to five systems. Every alert. Every time. That's not a workflow — that's an obstacle course."

Before & after — the analyst workflow

Research with 30+ analysts surfaced a consistent pattern of compounding friction. Here's what it looked like — and what the redesigned experience aims to deliver.

Current state — broken
Alert → 4–5 systems to disposition
01
Alert appears in legacy surveillance system — slow, outdated interface
02
Switch to system 2 to pull trade data — separate login, different layout
03
Switch to system 3 to cross-reference communications data
04
Missing data — gaps force judgment calls without full context
05
Over-document disposition to protect against QA scrutiny — defensive, not analytical
Future state — designed
Alert → unified platform, AI-augmented
01
Alert surfaces in unified platform — all relevant context pre-loaded
02
Trade & comms data consolidated — no system switching required
03
AI augmentation flags patterns, surfaces context, supports analyst judgment
04
Complete data available — analyst focuses on analysis, not workarounds
05
Disposition recorded with system-supported context — QA legible by design

The design direction

Consolidation. Bring the data analysts need into one place. Trust the analyst. Remove the defensive documentation burden. AI as augmentation. Help analysts move through higher alert volumes with greater confidence — keeping the human in the decision seat.

Core principle
"The goal was never to automate surveillance. It was to give analysts the tools to do surveillance well."

Outcome

The research and design foundation established a clear, evidence-based direction for a platform that can actually serve the people who use it. Analysts who participated consistently described the experience as the first time anyone had asked them what they needed. That itself is part of the outcome.

Read next
Fixing what happens after a patient leaves the hospital
A different industry, the same instinct — go in, find the humans, surface what's broken. Healthcare patient experience research at Baylor Scott & White.
Next case study
Go deeper
Work samples from this engagement
Deliverables from this initiative are being prepared for the work samples section. In the meantime, explore samples from related federal and banking engagements.
View work samples
🔒 Password protected
Next case study
03 — Baylor Scott & White
Healthcare · Patient Experience
Fixing what happens after a patient leaves the hospital
12–18 clinical stakeholder interviews. A research-led redesign of the post-discharge experience.