The Problem

Pharmaceutical brand teams spend their days trying to understand a complex web of relationships: Which payers cover their drugs? How is coverage changing? What's driving those changes?

This quote crystallized the problem. We weren't lacking data, we were lacking intelligence. Users could see that lives had changed, but understanding the driver required manual investigation.

Information Architecture

I developed an information architecture designed for expandability, reducing time to market for new offerings while implementing pervasive patterns that ease user adoption through familiarity.

Data Hierarchy

The key insight was that users thought in terms of questions, not data tables. I restructured the platform around four "data buckets" that map to user mental models:

AI Touchpoints System

Rather than building a separate AI chatbot, I embedded intelligence directly into the interface through hover states, click interactions, and macro-level summaries.

AI Insight Framework

I developed a comprehensive framework to classify and prioritize AI-generated insights. Every insight falls into one of four styles: Contextual Comparison, Trend + Driver Analysis, Actionable Recommendations, or Anomaly Detection.

Research Validation

Levels of Insight Design

I designed "AI Touchpoints" as contextual insights that appear exactly where users need them, at three levels of progressive detail:

The Solution

Unified Dashboard Experience

The new Payer Landscape dashboard provides immediate situational awareness. Users see total insured lives (317M), largest markets, key metrics, and trend data, all in one view. The timeline chart shows lives by channel (Commercial, Health Exchange, Managed Medicaid, Medicare, State Medicaid) over quarters, letting users spot patterns without switching screens.

The Channels and Plan Types donut chart reveals distribution at a glance, while the Lives by Geography map highlights states with the most changed lives, Illinois, Virginia, California leading with 8% trends.

Payer Directory

The Payer Directory shows which organizations are driving market changes. The Lives Changes visualization displays aggregate net gain/loss by organization, with dots indicating volatility (stable, moderate, high). The "Top Organizations Driving Change" panel, powered by AI, explains what's behind the numbers: Centene Health (+15.7M, +12%), MediCare Solutions (+15.7M, +12%), and negative movers like WellSpring Health Group (-4.22M, -2.7%).

Users can expand any organization to see subsidiaries, states of operation, channel breakdown, and reach contacts, all without leaving the page.

Geographic Intelligence

Users can drill from national views to states to counties, with each level providing appropriately detailed data. The California view shows 25 million total insured lives with an interactive county map. Each county row displays plans, formularies, top formulary, top plan, channel breakdown, rank, risk indicator, and opportunity score.

The organization context panel (Centene Corporation shown) provides quick access to payer details without losing the geographic context, total lives (13.5M), rank (#4), and horizontal bar charts comparing market share against competitors.

Relationship Intelligence

The payer organization profiles reveal complex hierarchies and business relationships. Centene Corporation (ID: 12345) shows 28 subsidiaries, 4th ranking, 1.5M Rx Lives, 500K Mx Lives. The hierarchy tree displays child organizations (Wellcare Health Plans, Fidelis Care New York, Health Net, Sunshine State Health Plan) with their types and lives counts.

The Business Relationships section shows services this organization provides and uses, who processes claims, who manages specialty pharmacy, who handles prescription fulfillment. Key contacts are listed with roles, LinkedIn links, and verified email addresses.

Formulary Analysis

The Payer Offerings view shows formularies by channel (Commercial, Managed Medicaid, State Medicaid, Medicare, Health Exchange) with tier structure distribution. Users can see at a glance which channels have the most complex formulary designs and filter by open/closed status, template/custom, and rank.

Hovering over any formulary shows the quick details panel: BayCare Health System (ID: 12345), controlled entity, design type, number of plans, and total lives covered.

Enterprise Responsive Design

Healthcare analytics users work on everything from cramped hospital workstations to massive trading floor monitors. I designed a responsive system spanning 8 breakpoints from 750px (Small) to 3820px (5XL).

Results & Impact

Design Impact

Process Contributions

Future Scalability

The information architecture I created allows the platform to expand into new data domains without requiring users to learn new interaction patterns. Shared design elements between Landscape iQ and Searchlight reduced design debt and positioned the team for unified platform experiences.

The Problem

Searchlight helps pharmaceutical teams track policy changes across payers: formulary updates, prior authorization requirements, coverage decisions. Missing a policy change can mean millions in lost revenue or compliance issues.

System Architecture

I developed a system architecture that separates concerns into four distinct experiences, each answering a specific user question:

System Architecture

I defined a new conceptual model that separated concerns into four connected experiences:

1. Alerting & Triage , "What happened?"

Searchlight Alerts (NEW), Alert Manager (NEW), and Email Digest (NEW) help users understand what's changed without logging into the platform.

2. Investigation , "What changed?"

Document List Summary (NEW), Document Workspace, and AI Summary + Compare (NEW) let users drill into specific policy changes with AI-synthesized explanations.

3. Curation , "What do I need?"

My Workspace (NEW), Flag + Notes (NEW), and Export Briefing Packet (NEW) enable users to curate and share findings with stakeholders.

4. Intelligence , "What should I know?"

Assessment Alerts (NEW), Expected but Missing (NEW), and Pattern Detection (2027) provide proactive insights about market dynamics.

Entry points include Email Alert, Direct Login, and Platform Alerting. Outputs include Briefing Packet, CSV Export, and Shared Alerts.

Research & Discovery

I conducted focus groups with Solutions Consulting to understand how Searchlight shows up in real customer conversations: what resonates, where deals stall, and what to prioritize for the 2026 roadmap.

Research Questions

User Journeys

I mapped five primary user journeys with target completion times:

Current State Analysis

I conducted a comprehensive heuristic evaluation of six core screens. The core problem wasn't just UI; it was the conceptual model.

The Solution

Dashboard Landing

The new dashboard provides immediate situational awareness with KPI cards: Total Documents (90, ↑23%), Lives Impacted (14.2M, ↑2.1M new), Owned Therapies Changed (47, ↑3 new), and Therapies Removed (12, ↓2 less affecting 1.5M lives).

Key Insights (AI Generated) surface three priority cards: "Attention Required" (UnitedHealth restricting Aimovig access, 1.7M lives), "Trend Alert" (Coverage duration extending to plan year, 8.3M lives), and "Positive Development" (Actemra added to 6 new formularies, 3.1M lives).

The Documents List shows alert activity with spike indicators, top payers, document counts, and "Review Changes" / "Flag for Briefing" actions.

Alert-Centric Workflow

Drilling into an alert (GLP-1 Competitors) shows detailed analytics: 28 total docs, 11 with changes, 17 no changes, top payers (CVS 8, UHC 6, Aetna 5), and 4 effective in 7 days.

The Document Volume chart (90 days) reveals activity patterns over time. The Payer × Therapy Changes matrix shows change severity (Multiple/Single/Minor) across therapies (Mounjaro, Ozempic, Wegovy, Zepbound) and payers (CVS, UHC, Aetna, BCBS).

Documents list includes inline AI summaries: "Mounjaro: Required medical info for systemic GLP-1 updated. Ozempic: Coverage duration changed from 6 months to plan year."

AI-Powered Prioritization

The Priority Review Queue ranks documents by business impact: Critical Priority (2), High Priority (5), Medium Priority (4). Each row shows alert name, description, priority badge, date effective, and lives impacted.

The AI Idea panel detects browsing patterns: "You've viewed 5 UnitedHealth authorization policies this week. Would you like to get notified of future changes?" This proactive intelligence surfaces relevant insights without requiring manual alert configuration.

Critical and Significant Updates appear in-context, allowing users to "View Updates" without losing their place in the queue.

AI-Synthesized Document View

The document detail page transforms raw PDFs into actionable intelligence. The AI Summary explains: "Key Change: CVS has added a step therapy requirement for Mounjaro. Coverage Impact: This represents a tightening of access, estimated impact ~800K commercial lives. Effective Date: January 20, 2026. Competitive Context: This change favors Ozempic as the first-line therapy. UHC and Aetna have not yet adopted similar requirements."

"Ask Ella" provides conversational access to deeper analysis. Lives Impact by Payer shows horizontal bar charts by organization. The By Change Type donut breaks down 10,377 changes: Coverage Expanded (45%), PA Updated (20%), Access Restricted (16%), Minor Updates (16%).

The Therapies in Document table lists each therapy with change indicators (+/-/~) and AI-generated summaries with page citations: "Actemra: The required medical information for systemic sclerosis-associated interstitial lung disease was updated. (Page 3)"

Alert Workflow Redesign

I completely redesigned alert creation with a streamlined flow. The Edit Alert modal shows: Alert Name, Filters applied (Benefit: Pharmacy, Lives Type: Insured Lives, Channel: Commercial, States), Frequency toggle (Daily/Weekly/Monthly), and Email notifications with "Send test email" option.

The Document Activity chart (December 2025) shows document and critical priority volumes over time, helping users understand alert patterns before configuring notifications.

Key principle: alert creation is always explicit. Users toggle it ON, it's never auto-created. Weekly Brief and Team sharing are prominently accessible for collaborative workflows.

Enterprise Responsive Design

Like Landscape iQ, I designed Searchlight to work across a full range of screen sizes, from 5XL (3821px) down to xSm (383px) across 9 breakpoints. The Document Compare and Document Preview Drawer views each have detailed responsive specifications.

At larger sizes, side-by-side comparison with diff highlighting is possible. At smaller sizes, the interface collapses to stacked views with preserved functionality. This ensures pharmaceutical teams can access critical policy updates from any device.

Results & Impact

The Searchlight 2026 redesign delivers measurable improvements in user efficiency and positions the product for client renewals:

The redesign achieves design parity with Landscape iQ, creating a unified platform experience across Norstella's Commercial Platform products.

System Architecture

The Problem

DemandTec is the industry pioneer in AI-powered retail pricing, optimizing prices across millions of SKUs for grocery and CPG retailers. But there was a fundamental trust problem:

Discovery & Research

I led a comprehensive research initiative involving stakeholder interviews, competitive analysis, and journey mapping of existing workflows.

User Personas

Journey Mapping

The creation wizard was intuitive, but once scenarios were generated, users hit a wall of unexplained numbers. The journey revealed three key actors: System Updates, the Pricing Solution, and the User.

Competitive Analysis

I conducted a comprehensive competitive audit evaluating DemandTec against PromoManager, TradePro, PromoOptimize, and emerging players.

The Solution

I designed a progressive disclosure system that let users choose their depth of analysis. The solution focused on three key innovations:

1. Transparent Scenario Dashboard

The redesigned Scenario Summary provides immediate category health assessment: unit volume (2.18M), sales ($7.49M), gross margin ($2.38M), and a synthesized "Healthy" status. Users know instantly whether optimization is working.

The Price Change Summary donut chart shows distribution across SKU/Zones, 45.88% increases, with clear breakdowns by percentage band. The Product Analysis quadrant maps every product by price change vs. volume change, color-coded by behavior: staple products (low sensitivity), price-sensitive products, discount-sensitive products, and items requiring investigation.

2. Price Elasticity Visualization

The most transformative addition was the price elasticity bubble chart. Each product appears as a bubble sized by revenue impact, positioned by price change (x-axis) and units sold (y-axis). Color-coding reveals behavior patterns: green for staple products (inelastic), blue for price-sensitive items, teal for discount-sensitive products, and pink for items needing investigation.

This visualization answered the "why" question instantly. Users could see that Product SKU:77 was recommended for a 5% price increase because it fell clearly in the "staple products" cluster, high volume, low price sensitivity. The science became visible.

3. Workflow Status Transparency

I redesigned the scenario listing to surface workflow status at every stage. Users can filter by scenario type, time period, and status (Optimized, Approved, Editable, Error, Cancelled, Recalculating). Color-coded status badges provide immediate visual recognition.

The production status indicator shows which scenarios are live ("In production" with green dot) versus pending. This eliminated the confusion about which prices were actually in market.

Promotion Calendar

For promotion planning, I designed a Gantt-style calendar showing all active and planned promotions across time. Users can toggle between day, week, and month views. The top KPI bar shows aggregate metrics: $800,000 total revenue, -13% vs prior period, $528,324 revenue delta.

Each promotion bar is color-coded by status and shows key dates. Overlapping promotions are stacked, revealing potential conflicts. This view transformed promotion planning from a spreadsheet exercise into visual strategy.

Change History & Comparison

Every scenario maintains a complete audit trail. The History tab shows all changes with timestamps, actors, and action types (Optimization, Scenario updated, Data received, New cost files, Re-modeled). Users can trace exactly how a scenario evolved.

The Scenario Compare feature allows side-by-side evaluation of up to three scenarios. Metrics are aligned row-by-row: unit volume, equivalent volume, sales, repeat sales, average price, gross margin. Users can finally answer "what would happen if we optimized for volume instead of profit?"

Exception Handling

One of the most requested features was better exception handling. I designed a warnings and errors panel that surfaces conflicts between AI recommendations and business rules. Each warning shows the rule violated, the affected products, and a "Click to debug" action.

For example: "No active Price Rule - SKU Limits was included in this scenario" appears as a warning with a direct link to resolution. This transformed exception handling from a manual audit into guided problem-solving.

Forecast Validation

To build long-term trust, I added forecast effectiveness tracking. The Forecast Details view shows actual vs. predicted performance over time, with the shaded confidence interval revealing prediction accuracy.

Top vendors for each period are ranked by contribution, and goals tracking shows performance against conversion targets across channels (Email, SMS, Mobile, WhatsApp). When users can see that the AI's predictions match reality, trust compounds over time.

Results & Impact

The redesign delivered measurable improvements across engagement, comprehension, and retention metrics. DemandTec was recognized as the SPARK Matrix™ leader in Intelligent Retail Pricing & Promotion Optimization.

The Problem

Marketing teams needed a more efficient way to segment their audience based on interactions with SMS campaigns to increase engagement and conversion rates. The existing process was manual and time-consuming, leading to inefficiencies and potential errors.

The workflow required exporting data to Excel, running formulas, and importing back into the campaign tool. Only technically savvy marketers could do it, creating bottlenecks and limiting campaign agility.

Objectives

• Simplify the process of creating and managing audience segments
• Enable marketers to quickly identify and target specific user behaviors
• Provide clear, actionable insights to optimize marketing campaigns

Research & Discovery

I conducted interviews with marketing professionals to understand their pain points and needs, performed competitive analysis of similar tools in the market, and reviewed user interaction data to understand common behaviors and patterns.

User Personas

Emily Johnson, Marketing Manager (Age 35) , Has been working in marketing for 10 years and manages multiple campaigns. She's tech-savvy and regularly uses analytics tools to track campaign performance. Her goal is to efficiently create targeted marketing segments to increase engagement and conversions. Key frustration: the current process is time-consuming, prone to errors, and lacks real-time feedback.

David Lee, Data Analyst (Age 28) , Has been working as a data analyst for 5 years, supporting the marketing team by providing insights and reports. He's proficient with analytics and data visualization tools. His goal is to provide accurate and actionable insights to the marketing team. Key frustration: the current segmentation tool lacks advanced analytics capabilities, and manual data extraction is time-consuming.

Key Findings

The user research revealed that marketing managers, data analysts, and marketing associates all require a more efficient and streamlined process for creating and managing audience segments. Three critical needs emerged:

• Straightforward interface to create complex segmentation rules
• Real-time feedback and clear visualization of segments essential for effective targeting
• Integration with existing marketing tools necessary to streamline workflows

Competitive Analysis

I evaluated Acoustic's segmentation capabilities against Salesforce Marketing Cloud, HubSpot, Klaviyo, Braze, and Mailchimp to identify gaps and opportunities.

UX Design Process

I explored several approaches through prototyping and user testing: query builders, natural language input, template libraries, and visual rule composition. I converged on a hybrid approach, a visual rule builder for power users with saved templates for common use cases.

Product Requirements

Based on the research, I identified three essential features:

Rule Composition: A visual interface for building complex segmentation logic with AND/OR operators, multiple filter types, and time frame constraints.

Real-time Segment Visualization: Immediate feedback showing audience size as users build rules, with demographic breakdowns and engagement metrics.

Integration with SMS Campaigns: Direct connection to all collected data from SMS, email, WhatsApp, mobile push, web forms, and journey touchpoints.

User Flow Documentation

I documented detailed user flows for three primary scenarios:

Flow 1: User sends any message from any journey from any account at any time , showing the path from segment list through rule composition to saved segment.

Flow 2: User enters and edits their message from a Journey "Journey ABC" where journey event's date is in the last 3 months , demonstrating the time-based filtering capabilities.

Flow 3: User first receives a message from the following: "Fragrance quiz", "Holiday greetings", "Snow Business Survey" in the last 6 months , showing multi-condition rule building.

The Solution

Template-Based Creation

Users begin by choosing between Dynamic or Static segments. Dynamic segments are subsets of users with attributes or behaviors that are updated in real time, allowing campaigns to target the most relevant and up-to-date data. Static segments are groups of contacts that meet certain criteria at a specific point in time and do not automatically update.

This choice happens immediately after clicking "Create segment," giving users clear mental models from the start.

Visual Rule Composition

The Rule Composer provides a two-panel interface: Filters on the left (Audience attributes and Events), Rules in the center (with AND/OR logic), and Reach on the right (real-time contact count).

Users can filter by audience attributes (subscription groups, frequency) and event types (Email, SMS, WhatsApp, Mobile Push, Journey, Web Form, Web Application). Each rule can include time frame constraints with precise date ranges.

The real-time reach calculation shows exactly how many contacts match the current rules, 140,000 in this example, updating immediately as users modify conditions.

Segment Management

The segments listing page provides a clear overview of all created segments with status indicators (Completed, Uncompleted, Corrupted), tags for categorization (US, Poland, Loyalty, Holiday, Customer retention, etc.), and creation/modification timestamps.

Users can search by name, sort by any column, and quickly identify segments that need attention. The color-coded status badges provide immediate visual feedback about segment health.

Quick Preview Panel

Rather than forcing users to navigate to a separate details page, I designed a kickout panel that slides in from the right. This allows users to preview segment details, including reach (114,000 contacts), rules, description, and notes, without losing context of the list view.

The panel shows different states: segment overview, rule details, and edit mode. Users can make quick changes directly from the panel or navigate to the full details page for comprehensive editing.

Segment Analytics

The full segment details page provides comprehensive analytics: engagement rate trends over time, demographic breakdowns, and performance comparisons. The "Cart abandonment" segment shows 97.5% engagement, 2.1% conversion rate, and $456,500 revenue attribution.

Users can view the complete contact list, export data, and access the rule builder directly. The overview tab provides quick stats while the contacts and rules tabs allow deeper exploration.

Results & Impact

The segmentation builder transformed how marketing teams worked with audience data, bringing what was previously a technical task into the hands of campaign managers.

The related AI Segments Builder project won 1st place at Acoustic's 2022 internal hackathon, demonstrating the team's innovation in making complex segmentation accessible to non-technical users.

Experience

Education

Skills

Tools

Recognition

Recommendations