Back
Dashboard Analytics UI Library
Dashboard Analytics UI Library
When I started the analytics redesign at Leena AI, the first thing I noticed was how scattered everything felt. Ticket volumes were in one place, agent performance somewhere else, bot metrics on a different screen… and users were constantly juggling between dashboards just to answer basic questions like “How are we really doing?”
That got me thinking: instead of making people piece insights together, what if we gave them a unified dashboard — one that shows the big picture at a glance but still lets you dive deep when you need details?
As the Lead Product Designer, I owned the entire process — research, wireframing, design, testing, and handoff. This case study is the story of how I turned fragmented dashboards into a single, cohesive analytics experience that helps HR leaders, managers, and analysts make decisions faster and with more confidence.
When I started the analytics redesign at Leena AI, the first thing I noticed was how scattered everything felt. Ticket volumes were in one place, agent performance somewhere else, bot metrics on a different screen… and users were constantly juggling between dashboards just to answer basic questions like “How are we really doing?”
That got me thinking: instead of making people piece insights together, what if we gave them a unified dashboard — one that shows the big picture at a glance but still lets you dive deep when you need details?
As the Lead Product Designer, I owned the entire process — research, wireframing, design, testing, and handoff. This case study is the story of how I turned fragmented dashboards into a single, cohesive analytics experience that helps HR leaders, managers, and analysts make decisions faster and with more confidence.
ROLE
Lead Product Designer
TEAM
1 Designer, 1 PM, 2 Engineers
DURATION
2 Months
RESULTS
Time-to-insight dropped from ~5 days (engineering-dependent exports) to real-time access
SLA reporting moved from monthly and manual to weekly and data-driven.
Feature request prioritization shifted from a Reactive to Data-driven approach.
why this mattered
Helpdesk tickets and product analytics showed users spending significant time just locating the right view before they could start analyzing anything.
Usage data showed people dropping off mid-flow — giving up before reaching the metric they actually needed.
Helpdesk tickets and product analytics showed users spending significant time just locating the right view before they could start analyzing anything.
Usage data showed people dropping off mid-flow — giving up before reaching the metric they actually needed.
The Problem
HR leaders, support ops, and analysts all needed the same underlying data for different jobs — but the existing dashboards gave everyone the same flat list of charts, with no hierarchy and no way to tell if a number was good or bad.
"I just want one place to see everything."
"Each module feels like its own world."
HR leaders, support ops, and analysts all needed the same underlying data for different jobs — but the existing dashboards gave everyone the same flat list of charts, with no hierarchy and no way to tell if a number was good or bad.
"I just want one place to see everything."
"Each module feels like its own world."
Research Summary
Affinity-mapped interviews and an insight-breakdown exercise surfaced four recurring complaints: no hierarchy (everything competed equally for attention), disconnected navigation across modules, filters that didn't persist, and metrics shown with no benchmark to judge them against.
A competitive review of Culture Amp, Tableau, and Pendo confirmed widget-based, drill-down layouts were already the category norm — the gap wasn't format, it was structure and context.
Affinity-mapped interviews and an insight-breakdown exercise surfaced four recurring complaints: no hierarchy (everything competed equally for attention), disconnected navigation across modules, filters that didn't persist, and metrics shown with no benchmark to judge them against.
A competitive review of Culture Amp, Tableau, and Pendo confirmed widget-based, drill-down layouts were already the category norm — the gap wasn't format, it was structure and context.
From Insights to Interface
Multi-level structure — top-line KPIs surfaced first, detail available on
drill-down instead of all at once.Persistent, comparative filters — team-vs-company benchmarks built into the
filter bar itself, not buried in a separate view.One visual language across 8 dashboard types (Overview, User Metrics, Bot
Metrics, Sessions, Queries, Ticket Volume, CM Performance, Agent Performance) —
shared chart library, color logic, and grid system, so switching modules never
felt like switching products.
Multi-level structure — top-line KPIs surfaced first, detail available on
drill-down instead of all at once.Persistent, comparative filters — team-vs-company benchmarks built into the
filter bar itself, not buried in a separate view.One visual language across 8 dashboard types (Overview, User Metrics, Bot
Metrics, Sessions, Queries, Ticket Volume, CM Performance, Agent Performance) —
shared chart library, color logic, and grid system, so switching modules never
felt like switching products.
Results & Impact
3x adoption across HR, IT, Procurement, Sales, and Finance, on a single design system spanning web, mobile, Slack, and MS Teams
2x user retention over the previous product experience
Full design ownership across three product generations:
chatbot MVP → generative assistant → agentic colleagueSeamless cross-platform experience with one system, not five separate ones
3x adoption across HR, IT, Procurement, Sales, and Finance, on a single design system spanning web, mobile, Slack, and MS Teams
2x user retention over the previous product experience
Full design ownership across three product generations:
chatbot MVP → generative assistant → agentic colleagueSeamless cross-platform experience with one system, not five separate ones
Retrospective
One size doesn't fit all — HR and Support Ops needed different KPIs, but still needed to feel like one product. Comparative views did more for adoption than any
single chart: once people could benchmark against a peer team or last month, the data stopped being just numbers and started being a decision.
One size doesn't fit all — HR and Support Ops needed different KPIs, but still needed to feel like one product. Comparative views did more for adoption than any
single chart: once people could benchmark against a peer team or last month, the data stopped being just numbers and started being a decision.
