Working with a major European insurance company, I led the design of a transformative AI-powered claims system. Over 4 weeks, our team designed two pilot workflows: a 2027 automated simple claims experience and a 2030 modular AI orchestration for complex cases, creating the strategic foundation that secured executive approval and budget for development.
Client
NDA - Major European Insurance Company
role
Lead Product Designer
timeline
4 weeks | 2025
team
1 PM, 1 Business Designer, 2 Product Designers
The problem
Claims processing is trapped between customer expectations and operational reality.
Modern customers expect Amazon-level service, but insurance claims still require manual document processing, phone calls, and weeks of silence. Through discovery sessions and exploratory research, we identified a critical gap: customers could track a $15 pizza delivery in real-time but waited weeks with zero visibility on thousand-dollar claims.
Current state pain points
Customer Experience: Manual FNOL (First Notice of Loss) submission, zero status updates, multi-week wait times for simple claims
Internal Operations: Significant manual processing, adjusters are buried in administrative work, and fragmented workflows across legacy systems
Business Impact: Declining customer satisfaction, rising operational costs, and competitive disadvantage
My role & approach
As Lead Product Designer, I directed the end-to-end design process:
Strategic Foundation
Participated in discovery sessions and expert interviews
Conducted exploratory research on AI trends and competitive landscape
Contributed to vulnerability analysis led by a business designer
Design Leadership
Led design team (2 Product Designers) through future scenario development
Designed two future-state pilot blueprints: 2027 simple claims automation and 2030 complex claims orchestration
Created executive storyboards and video simulations demonstrating the AI transformation vision
Research & discovery
Given the 4-week timeline, we focused on rapid, high-impact research:
Week 1-2: Understanding Current State
Discovery sessions with claims management teams
Exploratory research on industry trends, competitors, and emerging AI capabilities
Week 2-3: Validating AI Opportunities
Expert interviews with internal and external AI specialists
Vulnerability analysis evaluating business against AI disruption
Week 3-4: Defining Future Scenarios
Future scenario workshop mapping claims evolution (2025-2030)
Consolidation of most plausible scenarios
Key Research Insights
Adjusters spend the majority of their time on administrative tasks instead of expertise application
Digital-native insurers processing simple claims in under 24 hours
AI can handle significant document processing, but human expertise is critical for complex decisions
Market expectation shifting toward immediate resolution for simple claims
Design Strategy
Progressive Automation: 2027 to 2030
2027 Pilot: Automated Simple Claims
Target: Routine property damage with clear liability
Goal: Under 60 second resolution from submission to settlement
Approach: AI handles entire workflow, humans monitor for exceptions
2030 Vision: AI Orchestration for Complex Claims
Target: Multi-party liability, injury claims, catastrophic events
Goal: AI coordinates all routine tasks, humans focus on decisions
Approach: Modular agents handling specific tasks, human oversight on strategy
Key design decisions
Decision #1: Two-Phase Pilot Approach
Designing distinct 2027 (simple) and 2030 (complex) workflows allowed us to prove value incrementally rather than attempting risky full transformation immediately.
Decision #2: Conversational FNOL Instead of Forms
Natural language AI can extract structured data from conversation, reducing customer friction from multiple form fields to natural dialogue while improving data quality.
Decision #3: Real-Time Visual Damage Assessment
Visual AI analyzing photos instantly transforms multi-day assessment waits into 60-second processes for simple property damage.
Decision #4: Proactive Status Updates
Showing real-time claim progress eliminates repetitive status inquiry calls, improving customer satisfaction while freeing staff for complex cases.
The solution
Two pilot workflows: phased AI transformation
I designed strategic blueprints mapping future-state experiences, system architecture, and service touchpoints, creating the foundation for technical development.
Pilot 1: 2027 Automated Simple Claims
1. Instant FNOL via Mobile
Conversational interface extracts claim details from natural language
No forms, minimal friction
2. Visual Damage Assessment
AI-guided photo capture with computer vision analysis
Instant repair estimate with fraud detection
3. Automated Settlement
System validates coverage and auto-approves claims meeting thresholds
Instant notification with payment timeline
Total Time: Under 60 seconds from submission to approval
What I Created
Service Blueprint: Customer journey across all touchpoints, AI agent responsibilities, human oversight triggers, system integration requirements
Experience Concepts: Wireframes showing potential conversation flow, photo guidance, status transparency, approval patterns, adjuster monitoring dashboard
System Architecture: AI agent orchestration logic, data flows, integration points, exception handling processes
Key Innovation: Defined how AI could handle entire simple claims workflow while maintaining human oversight for edge cases.
Pilot 2: 2030 Complex Claims Orchestration Blueprint
1. Intelligent Triage
AI analyzes complexity and triggers senior adjuster assignment
Human oversight was established immediately
2. Automated Evidence Assembly
AI orchestrates photo collection, police reports, and medical communications
Delivers a comprehensive case package to the adjuster in hours vs. days
3. AI-Assisted Decision Support
System suggests liability analysis, flags conflicts, and surfaces precedents
Adjuster makes final determination with AI insights
4. Ongoing Orchestration
AI manages routine communications and tracking
An adjuster focuses on negotiation, empathy, and strategic decisions
What I Created
Service Blueprint: End-to-end orchestration across all parties, AI coordination logic, human decision points, workflow from triage to settlement
Experience Concepts: Adjuster dashboard wireframes, communication hub, evidence assembly interfaces, decision support visualizations, timeline showing AI-handled vs. human-decided activities
System Architecture: Multi-agent orchestration model, human-AI collaboration patterns, escalation mechanisms, quality assurance checkpoints
Key Innovation: Defined how AI agents coordinate complex workflows while preserving human expertise for judgment and strategy
System architecture overview
Five core AI agent concepts:
Claims Concierge: Customer conversation, data extraction, routing
Visual Assessment: Photo analysis, damage estimation, fraud detection
Triage: Complexity classification, routing, priority assignment
Orchestration: Multi-party coordination, evidence assembly
Settlement: Coverage validation, payment processing
Human Oversight: Exception monitoring, quality assurance, escalation triggers, strategic decision checkpoints
Results & key learnings
Project outcome
Our comprehensive deliverables (executive storyboard, service blueprints, and video simulations) enabled the internal claims team to successfully present the AI transformation vision to executive leadership in 2025, securing budget approval for development.
What I've learned
Blueprint work drives strategic alignment
Designing future-state systems without final UI forced us to focus on fundamental experience logic and architecture rather than pixel-perfect execution. This strategic approach (combined with video simulations that made the vision tangible) proved more valuable for executive decision-making than polished mockups would have been.
Modular thinking enables buy-in
Designing composable AI agents (vs a monolithic system) allowed stakeholders to envision a gradual rollout. "We can start with just a visual assessment," made the investment feel manageable and reduced perceived risk.
AI as partner, not replacement
Framing AI as intelligent support—not replacement—drove stakeholder embrace. When adjusters understood AI would handle the administrative burden so they could focus on complex decisions, the vision shifted from threatening to empowering.
What I'd Do Differently
Test with more end users
Limited direct customer and adjuster involvement. More co-design sessions would have strengthened experience logic and surfaced edge cases earlier.
Create Interactive Prototypes
Video simulations were effective, but interactive prototypes would have allowed stakeholders to dynamically explore "what if" scenarios during presentations.
Document More Edge Cases
Focus on primary scenarios left technical teams needing more guidance on edge cases like fraud, disputes, denials, and system failures.
Key Takeaway
Strategic design work isn't about polished pixels; it's about creating compelling visions that align stakeholders around a transformative future while defining the architecture to get there.
Success resulted from making a complex AI future feel real and reachable through clear plans, engaging storytelling, and modular thinking that let stakeholders picture the path from the current state to the future.