Sai Anjan - UX Designer

Context & Challenge

The Enterprise Challenge

A Fortune 500 technology company needed to modernize their internal analytics platform, which served over 5,000 employees across multiple departments. The existing dashboard was built in 2018 and struggled with:

Complex navigation requiring 5-7 clicks to access critical data
Static reports that required manual refresh and export
High learning curve preventing non-technical users from accessing insights
No intelligent recommendations or proactive insights
Fragmented workflows requiring users to switch between multiple tools

User Needs & Pain Points

Through extensive user research with 45 stakeholders—including data analysts, product managers, executives, and operations teams—we identified key pain points:

"I spend 30 minutes every morning just finding the right reports"

— Product Manager, 8 years experience

"The dashboard doesn't tell me what I should be looking at. I have to know what to ask for."

— Operations Director

"I wish I could just ask questions in plain English instead of learning complex filters"

— Marketing Manager

The Opportunity

The rise of conversational AI, demonstrated by Microsoft Copilot and similar enterprise tools, presented an opportunity to fundamentally rethink how users interact with complex data. Instead of forcing users to learn the system, we could make the system learn from users—through natural language queries, intelligent automation, and proactive insights.

The challenge was balancing powerful enterprise functionality with intuitive, conversational interfaces that feel as natural as asking a colleague a question.

Role & Tools

My Role

As the Lead UX Designer, I owned the end-to-end design process from research to implementation. My responsibilities included:

Conducting user research and synthesizing insights into design requirements
Designing the conversational interface architecture and AI interaction patterns
Creating information architecture for complex data hierarchies
Prototyping and testing AI-powered features
Collaborating with AI/ML engineers to define technical requirements
Leading design reviews and stakeholder presentations

Design Tools & Methods

Design & Prototyping

• Figma - Interface design and high-fidelity prototypes
• Framer - Interactive conversational AI prototypes
• Miro - User journey mapping and ideation
• Principle - Micro-interaction animations

Research & Testing

• UserTesting - Remote user testing
• Maze - Prototype testing and analytics
• Hotjar - User behavior analysis
• Dovetail - Research synthesis

AI & Conversational Design Tools

Microsoft Copilot Studio

Used Copilot Studio to design and prototype conversational interfaces. This tool allowed us to create natural language query systems that understand context and provide intelligent responses—similar to how Microsoft Copilot works across Office 365 applications.

Vibe Coding & AI-Assisted Development

Leveraged AI coding assistants to rapidly prototype conversational UI components and test different interaction patterns. This accelerated our iteration cycle from weeks to days.

Natural Language Processing (NLP)

Collaborated with ML engineers to design training data requirements for intent recognition, entity extraction, and context understanding—ensuring the AI could handle enterprise-specific terminology and workflows.

Approach & Solution

Design Philosophy: Conversational First

Inspired by Microsoft's approach to conversational AI, we adopted a "conversational first" philosophy. Instead of building a traditional dashboard and adding AI features, we designed the entire experience around natural language interaction, with visualizations as supporting elements.

"The best interface is no interface—but when you need one, it should feel like talking to a knowledgeable colleague."

1. AI-Powered Conversational Interface

We integrated a Copilot-style conversational interface that allows users to ask questions in natural language:

User asks:

"Show me sales performance for Q4 compared to last year"

✓

System responds:

Generates interactive comparison chart + key insights: "Q4 sales increased 23% YoY, driven primarily by enterprise segment (+45%)"

The AI understands context from previous queries, remembers user preferences, and can handle follow-up questions like "What about the European market?" without requiring full context.

2. Intelligent Automation & Proactive Insights

Instead of requiring users to manually refresh data or set up alerts, the system proactively surfaces insights:

Smart Alerts: AI identifies anomalies and significant changes, notifying users with context: "Your conversion rate dropped 12% this week—here's what changed"
Automated Reports: AI generates weekly/monthly reports based on user's role and viewing patterns, delivered automatically
Predictive Insights: "Based on current trends, you're likely to miss Q4 targets by 8%—here are recommended actions"
Workflow Automation: Common tasks like data export, report generation, and dashboard customization are automated through AI

3. Enhanced Data Visualization

While conversational interfaces handle queries, we redesigned visualizations to be more intuitive and interactive:

Before

• Static charts requiring manual refresh
• Fixed time ranges
• No drill-down capabilities
• Generic visualizations

After

• Real-time updates with intelligent refresh
• AI-suggested time ranges based on context
• Interactive drill-downs with natural language
• Personalized visualizations per user role

4. Design Decisions Based on User Feedback

Throughout the design process, we conducted 12 rounds of user testing with 60+ participants. Key iterations included:

Iteration 1: Voice vs. Text Input

Initial testing showed 85% preferred text input in office environments. We kept voice as an optional feature for accessibility and mobile use.

Iteration 2: AI Confidence Indicators

Users wanted to know when AI was "confident" vs. "uncertain" about answers. We added visual indicators and always provided source data links.

Iteration 3: Conversation History

Users requested ability to save and revisit conversations. We added conversation threads and searchable history.

Impact & Results

Quantitative Metrics

45%

Reduction in task completion time

Users now find insights in 2.3 minutes vs. 4.2 minutes previously

85%

Adoption rate within 3 months

4,250+ active users out of 5,000 employees

70%

Users regularly use conversational queries

3,000+ users prefer natural language over traditional navigation

60%

Reduction in support tickets

From 120/week to 48/week, saving $50K annually

Productivity Improvements

Time Savings: Average user saves 2.5 hours per week through automation and faster data access
Decision Speed: Executives report making data-driven decisions 3x faster
Learning Curve: New users become productive in 1 day vs. 2 weeks with old system
Report Generation: Automated reports save 15 hours/week per department

Qualitative Feedback

"This is exactly what we needed. I can ask questions in plain English and get answers immediately. The AI actually understands our business context."

— VP of Product, 12 years at company

"The proactive insights are game-changing. Instead of me hunting for problems, the system tells me what I need to know. It's like having a data analyst on my team."

— Operations Director

"I was skeptical about AI, but this feels natural. It's not trying to replace my judgment—it's augmenting it. The conversational interface makes complex data accessible."

— Marketing Manager, non-technical background

Lessons: AI's Role in Enhancing UX

1. Transparency Builds Trust

Users need to understand how AI makes decisions. We always showed source data and confidence levels, which increased trust by 40% in user surveys.

2. Augmentation, Not Replacement

The most successful AI features enhanced human decision-making rather than replacing it. Users appreciated AI suggestions but always wanted final control.

3. Context is Everything

Conversational AI works best when it understands user context—role, previous queries, current task. This required careful design of context management systems.

4. Progressive Disclosure

We learned to start simple and reveal complexity gradually. New users begin with basic queries, then discover advanced features as they become comfortable.

Business Impact

ROI: $2.3M annual value through productivity gains and reduced support costs
User Satisfaction: Increased from 3.1/5 to 4.7/5 in quarterly surveys
Feature Adoption: 80% of users actively use AI-powered features weekly
Competitive Advantage: Company now uses this as a recruiting tool, showcasing innovation

Visuals & Links

Key Screenshots & Mockups

Conversational Workflow Diagram

User-agent interaction flow showing how natural language queries are processed and responded to.

SaaS Dashboard Wireframe

Wireframe showing data visualization components, filters, and conversational interface integration.

AI Automation Flow Diagram

Copilot Studio UX writer automation flow illustrating how AI automates content generation and workflow tasks.

UX Research Insights

Infographic and framework visuals summarizing user feedback and research findings.

Conversational Design Principles

Diagram and summary visual outlining key principles for designing conversational AI interfaces.

Process Documentation

User Journey Maps - Before/After workflows
Information Architecture - Data hierarchy and navigation structure
Conversational Flow Diagrams - AI interaction patterns
User Testing Results - 12 rounds of testing documentation
Design System - Component library and AI interaction guidelines

Interactive Prototypes

High-fidelity interactive prototypes demonstrating conversational interface, AI-powered insights, and automated workflows are available for review.

Contact me at saianjan.margani@gmail.com for access to prototypes and detailed design documentation.

Reflection

This project demonstrated the transformative power of conversational design in enterprise software. By applying principles similar to Microsoft Copilot—natural language interaction, intelligent automation, and proactive insights—we created a dashboard that feels less like software and more like a knowledgeable assistant.

The success of this project validates that conversational AI isn't just a trend—it's the future of how we'll interact with complex enterprise systems. The key is designing AI that augments human intelligence rather than replacing it, always maintaining transparency and user control.