COMPANY
Shopify
ROLE
Lead Product Designer
Team
PM · Engineers · Content · Data
Timeline
5 Months
The Challenge
Only 2% of replatforming merchants activated within six months.
High-value merchants received dedicated onboarding specialists and personalized migration plans. Most merchants instead received generic onboarding flows that weren’t designed for migration complexity.
The Opportunity
Could AI replicate parts of the migration specialist workflow at scale?
The Outcome
+16.5% lead-to-customer conversion.
The exploration informed later iterations of Shopify’s AI onboarding architecture and guidance systems.
Understanding the Migration Problem
Replatforming merchants weren’t starting from scratch.
These businesses were migrating products, orders, integrations, workflows, and operational processes accumulated over years.
Many migrations stretched across weeks or months, with merchants often losing momentum before launch.
Existing onboarding wasn’t designed for this level of complexity, creating an opportunity for more contextual, adaptive guidance.
Understanding Migration Complexity
Research and discovery
To understand how onboarding specialists guided merchants through migration, I synthesized:
- consultant interviews
- onboarding workflows
- migration planning artifacts
- engineering and platform constraints
Several patterns emerged:
Replatformers needed reassurance, not just instructions
Migration involved uncertainty, unfamiliar terminology, and long timelines.
Migration workflows were highly non-linear.
Merchants frequently jumped between tasks depending on integrations, priorities, and blockers.
Context increased trust
Recommendations felt more credible when tied directly to a merchant’s specific migration setup.
Designing the System
Early explorations revealed that conversational AI alone wasn’t enough for long, complex migrations.
Merchants needed:
- structured onboarding
- persistent progress
- contextual guidance
- visibility into next steps
- flexibility across workflows
This shifted the challenge from designing a chatbot to designing a scalable onboarding system.
The experience combined conversational guidance with structured onboarding flows, allowing merchants to move between tasks without losing context.
Conversation vs Structure
We explored how conversational AI and persistent onboarding systems could work together without fragmenting the experience.
Task Grouping vs Overwhelm
Complex merchants needed visibility into migration scope without being overwhelmed by dozens of onboarding tasks.
Persistence Across Sessions
Migration onboarding often stretched across weeks or months. Guidance needed to persist beyond chat sessions so merchants could maintain momentum.
Designing AI Guidance Behavior
The onboarding experience needed to feel structured, contextual, and adaptive; not like an open-ended chatbot.
To guide merchants through complex migrations, I designed AI behaviors that:
- gathered migration context conversationally
- adapted recommendations based on integrations and business needs
- generated personalized onboarding guidance
- persisted onboarding progress across sessions
- surfaced next steps based on migration stage and blockers
The system balanced flexibility with structure, helping merchants feel guided without overwhelming them.
Exploration & Iteration
We explored several approaches to balancing conversational flexibility with structured onboarding guidance.
Early concepts included:
- chat-first onboarding
- nested task systems
- persistent guides
- grouped milestone architectures
Over time, the experience evolved from a primarily conversational assistant into a hybrid onboarding system combining AI guidance with persistent task visibility.
The Final Experience
The final onboarding experience combined conversational AI with structured onboarding guidance.
Rather than relying on open-ended chat alone, the system helped merchants navigate migration tasks, maintain progress across sessions, and receive context-aware recommendations throughout onboarding.
Merchants could:
- receive migration-specific recommendations
- generate personalized launch guidance
- revisit onboarding progress across sessions
- continue complex migrations without restarting context
Conversational intake gathered migration context and business needs dynamically.
Personalized onboarding guidance adapted recommendations based on merchant workflows and integrations.
Persistent onboarding guidance helped merchants maintain momentum across longer migration timelines.
Impact & Reflection
Impact
+16.5% lead-to-customer conversion.
The exploration informed later iterations of Shopify’s AI onboarding architecture and guidance systems.
What I Learned
Designing AI products requires designing both:
- the interface
- the conversation model
Designing AI products requires balancing flexibility with structure.
The conversation itself wasn’t the product, the real challenge was designing systems that helped merchants navigate complex workflows over time.