Back
Employee Feedback to Actionable Insights
Employee Feedback to Actionable Insights
Redesigning Leena AI's Employee Engage survey product
Redesigning Leena AI's Employee Engage survey product
Employee Engage was underperforming high support ticket volume, constant client hand-holding. I led the end-to-end redesign: four survey types, four personas, and the full analytics and action-planning experience.
Employee Engage was underperforming high support ticket volume, constant client hand-holding. I led the end-to-end redesign: four survey types, four personas, and the full analytics and action-planning experience.
ROLE
Lead Product Designer
Lead Product Designer
TEAM
2 Designers, 1 PM & 2 engineers
2 Designers, 1 PM & 2 engineers
DURATION
5 months · Shipped to production
5 months · Shipped to production
RESULTS
72% reduction in CS support tickets across survey workflows
64% dashboard adoption among HRBPs
38% faster manager action-plan creation
72% reduction in CS support tickets across survey workflows
64% dashboard adoption among HRBPs
38% faster manager action-plan creation
Why this mattered
20+ companies, 1.5M+ employees active on the product but it generated more support tickets than value. Setting up a survey, reading a dashboard, getting a report to leadership: none of it was self-serve.
The CS team was drowning in tickets across every part of the product. That was the signal: this wasn't a bug backlog. It was a broken experience.
20+ companies, 1.5M+ employees active on the product but it generated more support tickets than value. Setting up a survey, reading a dashboard, getting a report to leadership: none of it was self-serve.
The CS team was drowning in tickets across every part of the product. That was the signal: this wasn't a bug backlog. It was a broken experience.
The Problem
Area | What was broken |
|---|---|
Survey creation | Backend-only flow every launch needed manual CS support |
Reports | Nothing shareable, exportable, or scheduled |
Analytics & IA | One flat dashboard for every role nobody trusted it |
Action planning | Zero product support for deciding what to do about the results |
This wasn't a UI polish problem. It was an information architecture and role-design problem four different jobs were being forced through one generic flow.
Area | What was broken |
|---|---|
Survey creation | Backend-only flow every launch needed manual CS support |
Reports | Nothing shareable, exportable, or scheduled |
Analytics & IA | One flat dashboard for every role nobody trusted it |
Action planning | Zero product support for deciding what to do about the results |
This wasn't a UI polish problem. It was an information architecture and role-design problem four different jobs were being forced through one generic flow.
Research Summary
Client meetups, new-customer calls, CS ticket clustering, and cross-persona affinity mapping — to separate what users said they wanted from what was actually driving ticket volume.
Persona | Core job | What research changed |
|---|---|---|
HR Admin / VP | Create & manage surveys | Needed self-serve, not documentation |
Manager | Understand team sentiment | Wanted a narrow team view, not an org-wide dashboard |
Leadership | Org-wide visibility | Needed exportable/schedulable, not another login |
Employee | Complete surveys, feel heard | Wanted to see their input change something |
The biggest reframe: I assumed one dashboard with more filters would work. Research showed each role needed its own view of the same data — not a shared view with more knobs.
Client meetups, new-customer calls, CS ticket clustering, and cross-persona affinity mapping — to separate what users said they wanted from what was actually driving ticket volume.
Persona | Core job | What research changed |
|---|---|---|
HR Admin / VP | Create & manage surveys | Needed self-serve, not documentation |
Manager | Understand team sentiment | Wanted a narrow team view, not an org-wide dashboard |
Leadership | Org-wide visibility | Needed exportable/schedulable, not another login |
Employee | Complete surveys, feel heard | Wanted to see their input change something |
The biggest reframe: I assumed one dashboard with more filters would work. Research showed each role needed its own view of the same data — not a shared view with more knobs.
How Research Shaped the Interface
Design Decision: Four Survey Types, Not One Generic Flow. Split the product into four purpose-built types Periodic, Lifecycle/Milestone, Onboarding, Exit each mapped to a distinct employee-lifecycle moment. This was the single decision that unlocked everything downstream: a generic flow flexible enough for all four would have recreated the original hand-holding problem.
Design Decision: Four Survey Types, Not One Generic Flow. Split the product into four purpose-built types Periodic, Lifecycle/Milestone, Onboarding, Exit each mapped to a distinct employee-lifecycle moment. This was the single decision that unlocked everything downstream: a generic flow flexible enough for all four would have recreated the original hand-holding problem.
1. Survey Setup (End-to-End Creation Flow)
Five sequential tabs (Questions → Schedule → Audience → Test → Launch) with per-section status. Tested against a single long page tabs won because users always knew what was blocking launch.
2. Analytics Dashboard (Role-Based Data Views)
Separate dashboards per persona, each surfacing only what that role needs to decide not one dashboard with more filters.
3. Engagement Reports (Shareable & Exportable)
Exportable PPT/PDF summary with auto-scheduled email delivery, so leadership gets updates without logging in.
5. Action Planning (From Insight to Action)
AI-generated action plans tied to specific themes, one-click creation, scoped to the manager level rather than org-wide.
Testing the Assumptions
Usability tested setup, analytics, and action planning with HRBPs and managers
A/B tested the conversational survey format against the linear original
Tracked CS tickets, dashboard adoption, and action-plan creation post-launch chosen because they map to the four original problems
Changed based on testing: simplified the launch checklist, rewrote empty states, made anonymous-survey indicators more visible, added progress indicators (drop-off measurably dropped).
Usability tested setup, analytics, and action planning with HRBPs and managers
A/B tested the conversational survey format against the linear original
Tracked CS tickets, dashboard adoption, and action-plan creation post-launch chosen because they map to the four original problems
Changed based on testing: simplified the launch checklist, rewrote empty states, made anonymous-survey indicators more visible, added progress indicators (drop-off measurably dropped).
Results & Impact
First full quarter post-launch vs. prior quarter baseline
Area | Improvement |
|---|---|
Survey Setup - Self Serve | 90% of the total Surveys |
CS Support Tickets Reduced (Survey Creation + Report and PDF Generation + Others) | 72% |
Dashboard adoption (HRBPs) | 64% |
Manager action-plan creation | 38% faster |
PDF exported | 1000+ |
Reports Generated | 100+ |
Action Plan Generated with AI | 10000+ |
First full quarter post-launch vs. prior quarter baseline
Area | Improvement |
|---|---|
Survey Setup - Self Serve | 90% of the total Surveys |
CS Support Tickets Reduced (Survey Creation + Report and PDF Generation + Others) | 72% |
Dashboard adoption (HRBPs) | 64% |
Manager action-plan creation | 38% faster |
PDF exported | 1000+ |
Reports Generated | 100+ |
Action Plan Generated with AI | 10000+ |
Retrospective
Splitting into four survey types instead of one generic flow was the unlock; every downstream call had an obvious answer after that. In hindsight, shipping the employee-facing experience first would have built momentum faster than designing all four personas in parallel.
Splitting into four survey types instead of one generic flow was the unlock; every downstream call had an obvious answer after that. In hindsight, shipping the employee-facing experience first would have built momentum faster than designing all four personas in parallel.
