Redesigning Dashboards for a Smoother Contact Centre solution Experience

✅ Deliverables and Impact

• Insight Presentation: Comprehensive findings with supporting user quotes and workflow examples
• Recommendation Matrix: Each insight mapped to specific feature opportunities and implementation priorities

Design Influence

"This is the most usable research I've received—it helps me know exactly what to redesign." —UX Designer

Multiple insights were directly adopted into the redesign strategy:

• Real-time filtering and monitoring capabilities for supervisors
• Bulk management functionality and privilege management for administrators
• Enhanced alert visibility and prioritization systems with AI-powered anomaly detection

Project brief

Over a 6-month internship, I led two interconnected projects: first, identifying opportunities for AI automation across supervisor and admin dashboards; second, running a usability study on a pre-built AI navigation assistant prototype for the admin dashboard. My work uncovered workflow inefficiencies, mapped them to actionable AI solutions, and provided usability feedback to guide future AI product development.

The Problem: When Tools Create More Work

"I spend more time navigating dashboards than actually managing my team." —Supervisor participant

During initial stakeholder conversations, our Customer Success team shared feedback from enterprise clients about dashboard frustrations. However, the feedback was fragmented—some clients wanted "more data," others requested "less complexity." The product team was preparing to revamp two core dashboards but lacked a clear understanding of actual user workflows and pain points.

The Challenge: With limited documentation of current usage patterns and scattered client feedback, we needed to build a user-centered foundation for the redesign. The stakes were significant—these dashboards are critical tools for supervisors managing agent performance and administrators handling operational configurations.

Project 1: Identifying AI Automation Opportunities

As the sole UX researcher on this project, I led the research end-to-end, working closely with product and design teams to ensure insights would directly inform redesign decisions.

Goal:

Map pain points in current supervisor and admin workflows, and identify where AI could meaningfully reduce effort or improve oversight.

Research Approach:

I chose qualitative methods to uncover the nuanced, contextual issues that usage metrics often miss:

• Scoped and designed the study in collaboration with product and design.
• Recruited and interviewed 10 enterprise users through customer-facing teams
• Conducted remote interviews and live workflow walkthroughs.
• Thematic analysis using Dovetail to identify patterns across user groups

Key Decisions:

• Semi-structured interviews: I opted for flexible conversation guides to encourage natural storytelling and allow for unexpected insights to emerge.
• Remote methodology: This approach maximized participant reach across time zones while minimizing effort for busy enterprise users.
• Adaptive questioning: Early interviews revealed that admin dashboard usage was sporadic and task-specific than initially assumed, so I adjusted my focus accordingly.

Key Insights & AI Opportunities:

• Hard to Spot Agents with Consistent Threshold Violations (Supervisor): Supervisors must manually track agent behaviors like long handling time or delayed disposition, often missing anomalies during multitasking.

  → AI Opportunity: Use AI to detect recurring behavioral patterns (e.g., agents consistently taking long breaks, skipping disposition) and surface them as insights or predictive alerts.

• Low Visibility of Critical Status (Supervisor): Highlighted rows or passive color indicators are often missed due to scrolling or multitasking.

  → AI Opportunity: AI-driven alert systems that adapt visibility based on urgency and user context.

• Manual Monitoring Burden (Supervisor): Supervisors have to repeatedly download reports to check call metrics, login, break, etc.

  → AI Opportunity: AI-powered dashboards that auto-refresh and prioritize relevant metrics with options for customisation.

• Partial QA coverage (Supervisor): Quality checks relied on random sampling, leading to missed high-risk calls.

  → AI Opportunity: AI-based smart sampling models that prioritize calls with higher likelihood of issues.

• No Bulk Agent Management (Admin): Admins had to map or unmap agents one at a time, which was time-intensive for large teams, also required manaual effort in deciding how to dividie the team between campaigns, based on avaialable resources.

  → AI Opportunity: AI-assisted batch mapping suggestions that predict optimal agent assignments based on historical data with the ability to perofrm this assignment in bulk

• Lack of Role-Based Access (Admin): Uniform permissions for all roles increased the risk of errors and information overload.

  → AI Opportunity: AI-suggested permission sets tailored to actual role behavior, ensuring relevance and security.

Project 2: Usability Study – Admin AI Navigation Assistant Prototype

Context:

The admin dashboard AI navigation assistant was a pre‑built prototype (created before my internship) intended to help admins quickly locate settings and complete task‑based actions through guided navigiation. Because it aligned closely with navigation challenges uncovered in interviews, I tested it immediately after each participant’s interview, while their workflow context was fresh.

Goal:

Evaluate the usability, clarity, and perceived value of the prototype to understand how well it addressed admins’ navigation needs and where improvements were necessary.

Research Approach:

• Embedded usability testing directly after the interview phase to preserve context.
• Used Maze for remote, task-based sessions.
• Gathered qualitative feedback on clarity, usefulness, and improvements.

Findings:

• Concept validated as valuable, but admin dashboard usage is infrequent and highly task-specific.
• AI was seen as more impactful for high-frequency workflows like monitoring and proactive alerts.
• Participants expected that the assistant could go beyond guidance and actually perform certain routine tasks on their behalf, reducing manual effort.

The Realization:

While participants appreciated the concept, AI assistance needed to focus on surfacing task-relevant actions, not broad navigation improvements. For sporadic, high-stakes usage, precision mattered more than convenience

🧭 Project timelines and collaboration

Months 1-2: Stakeholder alignment, research planning, participant recruitment, and prototype integration

Months 3-4: Conducted 10 user interviews with embedded usability testing

Month 5: Thematic analysis and insight synthesis using Dovetail

Month 6: Final presentation, design handoff, and documentation for future research

Collaborative Elements

Throughout the project, I worked closely with product managers and designers to:

• Refine research questions based on design priorities
• Provide interim insights to guide early wireframe development
• Create actionable documentation that spoke directly to implementation needs

Impact & Outcomes

• Delivered an insight-feature matrix linking pain points to AI opportunities.
• Provided usability feedback that shaped AI feature prioritization.
• Contributed to the ongoing redesign of both dashboards.

📘 What I learned

This project significantly developed my skills in enterprise research and cross-functional collaboration:

Enterprise User Recruitment: I learned to navigate complex organizational structures and leverage client-facing teams to access participants effectively.

Adaptive Research Methods: The ability to adjust my approach mid-study—such as refocusing on task-specific admin workflows—proved crucial for uncovering actionable insights.

Strategic Communication: Developing the insight-feature matrix taught me to translate research findings into formats that directly support design and product decisions.

Systemic Thinking: Following up on unexpected insights about the support portal demonstrated how comprehensive user listening can reveal cross-product opportunities.

🚀 What’s Next

If I could extend this research, I would:

Usage Segmentation: Analyze how dashboard workflows vary by role, team size, and organizational context to inform feature prioritization.

Longitudinal Study: Track how user workflows evolve as new features are implemented to measure impact and identify new optimization opportunities.

🎤 Final Takeaway

This project reinforced my belief in the power of qualitative research to uncover systemic issues. By focusing on user workflows and listening for the stories behind workarounds, I was able to identify friction points that had become so normalized that users barely mentioned them initially.

The experience taught me that effective UX research in enterprise contexts requires balancing thorough investigation with practical constraints, and that the most valuable insights often emerge from understanding not just what users do, but why they've learned to do it that way.

Most importantly, this project demonstrated how user-centered research can serve as a bridge between customer needs and product strategy, ensuring that design decisions are grounded in real user experiences rather than assumptions.