Overview

eBay Strategic Alignment
Using AI to improve how teams work enabling faster execution on Focus Categories, eBay Live, C2C, and global scale
RSM: Scalable Solutions Internal AI Workforce Fluency

My Dual Role: Product Manager and UX Lead, embedded in the People Team, working daily with IT Services & Solutions. Same team, both hats, no handoff.

Global employee experience across Benefits, Compensation, and critical employee journeys in 30+ global offices
Workday Help knowledge base led the company-wide initiative and continue evolving it through AI and automation
Business to technology translation turning employee and business needs into requirements IT and Engineering can build and scale
60-person team across engineering, HR COEs, legal, and Workday specialists, delivered in 14 months

Results at a Glance: Reduced HR case volume by 24%, increased employee self-service by 17%, deployed 3 distinct AI agents (self-service benefits, financial wellbeing, and HR case deflection), and delivered a modern, conversational AI-driven HR experience for 10,000+ global employees across APAC, EMEA, and AMER all within 14 months and within budget.





The Problem: High-Friction Workflows, Overwhelmed HR

eBay’s HR support was under serious strain. Employees across 30+ global offices struggled to find answers to even basic questions static knowledge base pages, disconnected portals, and fragmented benefits information left them no choice but to submit a support ticket. This created a cascading problem: HR case volume was unsustainably high, costs were rising, and employees felt unsupported.

My starting point was workflow analysis, not solution design. I analyzed 12 months of support tickets, mapped employee journey friction points across three separate HR portals, and interviewed 40+ employees and HR specialists to identify where the pain was highest and the AI opportunity was greatest. This research directly shaped the prioritization decisions I made as PM: three distinct workflow problems, three targeted AI agents.

The Three Highest-Value Workflow Opportunities
Agent 1: Self-Service Benefits
Most common HR tickets were benefits questions. An AI agent that gives employees instant, accurate answers without escalation.
Agent 2: Financial Wellbeing
Employees needed personalized guidance on 401(k), stock options, and financial resources high-anxiety questions that HR specialists were spending hours answering individually.
Agent 3: HR Case Deflection
A catch-all conversational agent covering PTO, tax, onboarding, and policy questions reducing human touchpoints on routine inquiries and freeing HR for complex cases.




How I Led This: PM + UX + Engineering Bridge

Bridging People Team and Engineering (IT Services & Solutions)

The People Team knew exactly what was painful. Engineering knew what was technically possible. Neither group naturally speaks the other's language, and that gap is where most AI projects stall. I lived in that gap translating in both directions and keeping both sides accountable to real outcomes, not just shipped features.

What bridging the gap looked like in practice
  • ▸Translated 40+ HR use cases into structured product requirements Engineering could scope and estimate
  • ▸Ran bi-weekly alignment sessions between HR COEs, IT Services & Solutions engineers, and Workday specialists
  • ▸Prioritized the backlog based on ticket volume, employee impact, and engineering feasibility not just stakeholder loudness
  • ▸Defined success measures before any build started adoption rate, deflection rate, employee satisfaction so we had a shared definition of done

Think: Identifying High-Impact Workflow Opportunities

  • Analyzed 12 months of support tickets (8,000+ cases) to identify root causes and frequency benefits questions were #1, financial wellbeing #2, PTO/tax/onboarding #3
  • Mapped employee journeys across three separate HR portals to reveal where friction was highest
  • Interviewed 40+ employees across AMER, APAC, and EMEA to understand workflow breakdowns in each region
  • Collaborated with HR COEs to identify knowledge gaps, compliance constraints, and what "instant, accurate answer" meant for each use case



Make: Designing the Agentic Workflow Solutions

I designed three distinct agentic AI workflows, each targeting a specific high-impact problem. The shared platform was Workday Assistant a conversational AI built into the tool employees already used but each agent had its own conversation architecture, content structure, fallback paths, and success metrics.

Agent 1: Self-Service Benefits
Designed conversational flows that answered benefits questions instantly health plan options, enrollment windows, dependent coverage, FSA/HSA rules without requiring an HR ticket. Structured content from 12 HR COEs into AI-retrievable knowledge. Built fallback paths that escalated gracefully when questions exceeded the agent's scope. Result: handles 17% more self-service benefits inquiries without human involvement.
Agent 2: Financial Wellbeing
Designed a personalized guidance agent covering 401(k) contribution questions, stock option vesting, financial resources, and EAP (Employee Assistance Program) access. This was the most sensitive workflow tone, privacy, and accuracy were non-negotiable design requirements. Collaborated with Legal and Benefits COE to ensure every response met compliance standards. Result: eliminated the need for individual specialist consultations on financial wellbeing basics.
Agent 3: HR Case Deflection
A broad catch-all agent covering PTO policies, tax information updates, onboarding steps, and policy FAQs. Designed with smart intent recognition and contextual routing if the agent couldn't answer with high confidence, it routed to the right specialist instead of leaving the employee stranded. Result: reduced overall HR case volume by 24%.
Cross-cutting design decisions:
  • Integrated an AI-powered knowledge base that surfaces the right content based on employee intent, not keyword matching
  • Designed context-aware answer flows that remembered conversation state employees didn't have to repeat themselves
  • Structured content with HR COEs so answers were accurate, compliant, and searchable by the AI engine
  • Built regional variations for APAC and EMEA different policies, languages, and compliance requirements required distinct conversation paths



Check: Validating the Experience

  • Conducted usability testing with real employees simulating HR tasks via the chatbot
  • Used behavior analytics post-launch to track search success rates, deflection rates, and engagement
  • Iterated on chatbot prompts, fallback paths, and tone to enhance the conversational UX




The power of AI & Workday Assistant

The game-changer? Workday Assistant - an AI chatbot that uses natural, conversational language to interact with employees just like a human HR rep would.

It answered questions like:
  • “How do I update my tax info?”
  • “When is open enrollment?”
  • “Where can I find the PTO policy?”

The assistant handled these interactions instantly, without any need to submit a ticket. This conversational approach not only made the experience more intuitive but also dramatically increased trust and engagement.

AI was central to this transformation. It powered:
  • Smart search that understood user intent
  • Auto-surfaced answers based on context
  • Predictive suggestions that reduced friction and increased speed to resolution

AI didn’t just improve the UX - it redefined how employees interact with HR at eBay.





What This Taught Me About Enterprise AI Transformation

This wasn’t primarily a UX project it was a workflow transformation project. The design was the visible layer. The real work was identifying the right workflows to target, securing organizational alignment, translating human needs into technical systems, and driving adoption at scale across a global workforce that had never used conversational AI for HR before.

  • Workflow analysis beats solution assumption: The three agents I built weren’t the first ideas on the table. They came from rigorous ticket analysis and employee research. Many teams skip this step and build AI solutions in search of a problem. I built solutions for proven, high-volume problems.
  • AI adoption is a change management challenge: Technical deployment was the easy part. Getting employees to trust the agents and HR leadership to believe deflection wouldn’t mean worse outcomes required months of communication, visible success data, and gradual rollout with feedback loops built in.
  • The PM-UX-Engineering triangle requires a translator: Sitting between People Team and IT Services & Solutions, I had to constantly translate in both directions. HR COEs needed to understand what AI could and couldn’t reliably answer. Engineers needed to understand why a 94% accurate agent was still too risky for financial wellbeing advice without a human escalation path.
  • Measure before you build and after you ship: I defined deflection rate, adoption rate, and employee satisfaction targets before a single agent was designed. Post-launch, those metrics drove every iteration. When Agent 3 showed lower-than-expected engagement in APAC, we used behavioral analytics to diagnose and fix the conversation flow within one sprint.




Impact

  • 17% increase in employee self-service employees now resolve benefits, financial, and policy questions without ever opening a support ticket
  • 24% reduction in HR support case volume freeing the HR team to focus on complex, high-value cases that genuinely need human judgment
  • 10,000+ employees served globally across AMER, APAC, and EMEA with regional conversation variations for local policy differences
  • 3 production AI agents deployed self-service benefits, financial wellbeing, and HR case deflection each with its own success metrics and continuous improvement loop
  • Significant operational cost reduction by eliminating human touchpoints on routine, high-volume tasks
  • Built a reusable agentic workflow framework adopted by other eBay internal teams for their own AI agent deployments

The technology part of this project was not the hard part. Changing how a global workforce interacts with HR earning trust, driving adoption, and sustaining it across 30+ offices took longer and required more care than any of the engineering work did.





Beyond the People Team

People Team AI Enablement

Shares AI best practices, tools, and workflows with the People Team through regular internal sessions upskilling HR staff on AI agents, prompting, and automation so the team can leverage AI independently in daily work.

Global Markets Success Workshop

Designed and facilitated an AI Transformation Workshop for eBay's Global Markets Success team introducing agentic workflows, GenAI tools, and practical adoption frameworks to drive AI literacy across the global commercial organization.



To discuss the detailed workflow analysis, agent architecture, change management playbooks, and adoption measurement frameworks behind this project or to explore how this experience maps to enterprise AI transformation at scale connect with me on LinkedIn / Nancy-UX.

On this page