Image
Image

AI Innovation Workshop

Role: UX Design Manager

I challenged 20+ cross-functional partners to think beyond optimized keyword-based queries and instead solve for complex, ambiguous or open ended search queries with emerging AI technologies.


Management Impact & Key Takeaways

Architecting Agentic Search

Established the foundational design framework for the Core AI team's future agentic search initiative.

Cross Functional Influence

Mobilized 20+ XFN leaders across disconnected organizational silos (Search, Ads, Structured Data, Verticals) under a unified AI vision.

Executive Buy-In

Successfully secured executive alignment on next steps, user testing, and technical feasibility reviews.

Design Strategy Playbook

Created a repeatable blueprint for facilitating rapid, high-impact AI vision jams across complex e-commerce platforms.

Customer Problems

Through pre-jam research synthesis and discovery, we framed five core pain points:

Expertise
Barrier

Specific searches require tips, tricks, and domain knowledge. Searching feels like a complex puzzle rather than an intuitive process.

Unaware of
Options

Buyers wanting to complete a project or collection struggle to find personalized recommendations on where to start.

Overwhelming Results

Broad searches lead to endless, difficult-to-parse Search Results Pages (SRP).

Contextual Intent Gaps

Scanning item details across multiple listing pages to find basic criteria (e.g., shipping speed, exact fit) is frustrating.

Cognitive Load & Tab Overload

Comparing multiple listings forces users to toggle continuously between SRP and item pages.

Leadership & XFN Sprint Architecture

As Strategic Lead, I structured a fast-paced, 5-day design jam methodology to bridge the gap between AI technical potential and real customer value:

Day 1

Kickoff

Aligned on domain walk-throughs, UX literature, search ads, and technical AI guidelines from leadership.

Day 2

Opportunity Framing

Synthesized cross-functional ideas into targeted problem statements and initiated rapid sketching.

Day 3-4

Co-Design & Prototype Iteration

Presented consolidated concepts to cross-functional stakeholders and PM leadership.

Day 5

XFN Share Out

Facilitated concept design across breakout groups, refining raw ideas into interactive concept flows.

Day 5

Post-Sprint (Executive Alignment)

Led executive review with leadership (Design VP & Search Product VP) to align on vision, strategic fit, and implementation roadmaps.

Design Concepts & AI Paradigm Shifts

Design Concepts & AI Paradigm Shifts

I guided the team in moving from abstract AI capabilities to concrete interaction models:

Smart Filters

Problem Addressed:

Overwhelming search results

AI Solution & Interaction Pattern:
Dynamically surfaces contextually relevant filter chips powered by LLM general knowledge and user preferences (e.g., "Accommodates multiple children", "Front cargo bucket").

Global Agent Handoff

Problem Addressed:

Disconnect between high-level AI assistants and deep inventory browsing

AI Solution & Interaction Pattern:
Guides shoppers through initial query framing via global agent, then smoothly transitions them back to a tailored SRP for deep browsing.

Conversational Commerce

Problem Addressed:

Rigid keyword search & decision fatigue

AI Solution & Interaction Pattern:
Integrates an in-line natural language assistant (eBay.ai) on the search results page to let buyers refine results conversationally without breaking browse flow.

Adaptive Search Bar

Problem Addressed:

AI latency and rigid autocomplete

AI Solution & Interaction Pattern:
Detects natural language input early to expand search input real-estate, offers predictive conversational suggestions, and reduces perceived latency.

Smart Filters

Problem Addressed:

Overwhelming search results

AI Solution & Interaction Pattern:
Dynamically surfaces contextually relevant filter chips powered by LLM general knowledge and user preferences (e.g., "Accommodates multiple children", "Front cargo bucket").

Conversational Commerce

Problem Addressed:

Rigid keyword search & decision fatigue

AI Solution & Interaction Pattern:
Integrates an in-line natural language assistant (eBay.ai) on the search results page to let buyers refine results conversationally without breaking browse flow.

Global Agent Handoff

Problem Addressed:

Disconnect between high-level AI assistants and deep inventory browsing

AI Solution & Interaction Pattern:
Guides shoppers through initial query framing via global agent, then smoothly transitions them back to a tailored SRP for deep browsing.

Adaptive Search Bar

Problem Addressed:

AI latency and rigid autocomplete

AI Solution & Interaction Pattern:
Detects natural language input early to expand search input real-estate, offers predictive conversational suggestions, and reduces perceived latency.

Challenges & Pivots

The Challenge

Executive leadership pushed for purely AI-driven innovation, creating pressure to ship flashy, AI-for-AI’s-sake features without clear user value.

The Pivot

Rather than chasing novelty, I aligned the team around a strict value filter—prioritizing scalable AI applications tied directly to core user pain points and measurable business KPIs.

Prioritization Criteria

I established a clear framework to evaluate and select which initiatives to move forward.

Solves a customer problem that can't be adequately addressed with current tools

Gen AI and/or an LLM is uniquely suited to support solution

Cost
(Engineering + Capacity)

Reach and
business impact

Previous: Leadership Style

Next: Search Results Page Redesign

Logo