TL;DR
Designed an AI-driven exception review platform for Goldman Sachs' Controllers division, turning a 4–7 hour manual exception review into a minutes-long one. The work spanned multiple user groups within Controllers (Revenue, Risk, Regulatory Reporting), with additional input from adjacent Risk divisions (Liquidity Stress Testing, Credit Risk). Took the work from concept to MVP, owning the full design as the sole UX designer on the engagement. The RAG-triage information architecture I developed with the working group became the model the team carried into build.
The Problem
Goldman Sachs' Controllers sign off on the accuracy of financial datasets every day, month, and quarter. The existing tooling gave them raw exception data but no intelligence on top of it.
A Revenue controller investigating hundreds of data quality discrepancies ("exceptions") before PnL close worked the same way as a Risk controller clearing risk-metric breaks, or a Regulatory Reporting controller resolving disclosure issues before SEC and Fed filings: review each exception by hand, switch between systems to piece together context, document every action for audit. Roughly 70% of exceptions traced back to the same handful of root causes, but the tool couldn't see that, so each one went through the full manual cycle.
The opportunity wasn't to build another exception view. It was to rethink how a controller relates to the data: AI does the heavy lifting, the human reviews and signs off.
My Role
Sole UX designer on the engagement, owning the full process — discovery workshops, system mapping, wireframes, prototypes, and user testing facilitation. Worked directly with the Goldman Sachs Controllers AI Working Group in weekly design cycles, collaborating with their UX Lead, Working Group Lead, and domain SMEs across Revenue and Risk teams.
Discovery & Design Philosophy
Early work focused on building a mental model of the exception review lifecycle through stakeholder workshops and technical deep-dives across the domain.
The critical insight emerged early: stakeholders didn't want a chatbot. They wanted more proactive AI suggestions embedded in the workflow, not a reactive conversational interface. This became the foundational principle.
Key Decisions
Through rapid, stakeholder-validated iteration:
Before/after over rationale. Summary tables show "current → suggested value" for fast action. AI reasoning lives in hover tooltips and drill-down panels via progressive disclosure.
Grouped exceptions with bulk actions. Since ~70% are duplicates, grouping by pattern with per-action-type rows dramatically reduces cognitive load while preserving audit traceability.
Non-blocking AI processing. Replaced a full-screen blocking modal with row-level inline indicators and a collapsible side drawer — Controllers keep working while AI runs.
Rejection as a forward path. Rejecting a suggestion prompts context addition (file uploads, data source selection) and AI re-analysis, not a dead end.
Invisible audit trails. Every action logged automatically with timestamps and attribution. Zero extra documentation effort from the user.
Expanding to Risk
As Revenue matured, I designed a version-branching strategy extending the same architecture to Risk Controllers. Beyond a different status model (three-state: Not Resolved, Non-blocking, Blocking — where blocking exceptions prevent sign-off), Risk discovery surfaced a new capability: "risk explains" — AI-driven post-exception attribution analysis that helps Risk Controllers understand why a break occurred, not just how to resolve it. This opened up risk-specific action types like adjustments, rolls, and risk acceptances that don't exist in the Revenue workflow.
Current State
This is a WIP and the prototype is in structured user testing with with different user groups. The north star: turn a multi-hour daily review into minutes — AI does the heavy lifting, humans review and make the final call.
