Enterprise Search

Global Analytics Federated Search

Re-architected the search experience across a fragmented analytics ecosystem, connecting 40+ unique data sets into one workflow so analysts can discover, validate, and act without tool-hopping.

Enterprise SearchAnalytics PlatformsWorkflow AutomationRegulated Environments

Overview

Analysts needed to search across multiple disconnected tools and data sources. The experience was slow, inconsistent, and relied heavily on manual work, often requiring people to leave their workflow, reformat identifiers, and repeat searches tool-by-tool.

I led the product and UX direction for a federated search experience that unified discovery across 40+ unique data sets, standardized results and filtering patterns, and reduced workflow friction with normalization and automation patterns built into the UI.

The outcome was a single, enterprise-ready search workflow that established patterns other teams could reuse—improving findability, consistency, and user trust in results.

Project Details

Role

Product Manager + UX Director (Portfolio Lead)

Timeline

Multi-phase delivery (shipped on schedule)

Team

Cross-functional delivery with product + engineering (small core team, multi-team ecosystem)

Deliverables

Personas · Journey maps · IA + taxonomy patterns · Lo/Hi-fi prototypes · User testing · Design specs · Backlog/user stories

Tools

Figma / Adobe XD · Jira / Confluence · Heuristic evaluation · Prototyping + usability testing

Discovery

Shadowing & Workflow Mapping

Captured how users discover, confirm, compare, and export information across their daily workflows

Heuristic Evaluation

Evaluated current search workflows and identified key pain points and friction areas

Card Sorting & Labeling

Exercises to align filters and categories to user mental models and terminology

Prototype Testing

Validated findability, confidence, and speed-to-action through iterative testing cycles

Define

Problem Synthesis

Fragmented Discovery

Analysts had to search 40+ separate data sets manually, with no unified view across sources.

Manual Normalization

Converting identifiers between systems required spreadsheets and external tools, adding 30+ minutes per task.

Trust Gaps

Users couldn't verify data source or quality without deep system knowledge, reducing confidence in results.

User Workflow Loop

Search
Filter
Triage
Validate
Compare
Act

User Stories

Intelligence Analyst

"Search across all relevant data sources in one place"

Solution

Federated search connecting 40+ data sets with unified results

Research Lead

"Trust and verify where search results come from"

Solution

Source transparency and confidence cues in result cards

Operations Team

"Normalize identifiers without context-switching to spreadsheets"

Solution

In-workflow normalization reducing manual effort by 90%+

Design

User Workflow Loop

Search
Filter
Triage
Validate
Compare
Act

Federated Search Must Preserve

Source Transparency

Users need to know where results come from to trust them

Confidence Cues

Visual indicators help users assess result quality and relevance

Consistent Patterns

Standardized result formats reduce cognitive load across sources

In-Workflow Normalization

Normalize identifiers inside the workflow so users don't context-switch to spreadsheets/tools

Deliver

40+ Data Sets Connected

Connected discovery across 40+ unique data sets into one federated workflow

~30 min → ~2 min

Reduced manual normalization effort inside key workflows (example improvement for a repeated normalization task)

Reusable Enterprise Patterns

Established reusable enterprise patterns that other product teams could adopt for consistency