↑ Back to top

Overview

Problem & Solution

Contributions

Research

Achivements

Takeaways

Microsoft AI Insights

Revamped the customer engagement insights platform, boosted task efficiency by 33% for users across 5,000 businesses.

microsoft.com/dynamics-365/customer-service

Context

Contact centre products, powered by AI, enables businesses to effectively engage with their customers who are seeking assistance. I had the opportunity to work with the AI Insights product.

My work involved in transforming Insights Platform. The platform collects data from customers engaging with businesses across all channels. Analyzing data and improving LLM responses when a customer engages via multiple channels is critical to improving the models.

Role

I served as the primary designer and researcher across the project's lifespan. I was involved in end-to-end design process
‍

→ Interviewed Stakeholders
→ Facilitated focus group sessions
→ Mapped end-to-end user workflows
→ Wireframe A/B test
→ Designed interfaces using design system
→ Created prototypes for usability testing

Time Line
Sep 2022 - December 2022

Project Type
Enterprise Web App

Team
UX Researcher
Principle Designer
2 Developers
Technical Writer
Project Manager

Disclaimer 🔒

I cannot share the details here but I would be more than happy to chat about it alongside my overall experience over a call.

Problem

The platform gathers data on customers engaging with LLM. Without collecting and analyzing every conversation, it is impossible to know how the model performed. This is where the conversation analysts(users) have a significant challenge.

In the current platform, users analyze thousands of monthly AI-Customer conversations! Each conversation takes 30 minutes long!

Solution

A redesign implemented to modernize the platform and integrate new features tailored to user and client needs. This led to increased user satisfaction, improved efficiency in detecting LLM failures, and reduced customer resolution times.

Contribution in numbers

4

Stakeholders Interviewed
(Analyst, PM, CTO)

4

Competitive Audits

2

Focus Group Sessions
(Planned and facilitated)

12

User Engagements

1

Round of A/B Testing

3

Rounds of Iteration

8

New Components Introduced

Grounded in research driven approach

The project received positive results due to the data-driven approach to design. I tackled understanding the problem at an deeper level using the following methods/sources:
‍

Heuristic Evaluations

Competitive Audits

Focus Group Interview

Usability Testing

User and Client Feedback

Key achievements

I gathered diverse user feedback and aligned design objectives with business goals to create a polished final product. Through iterative design and data-driven decisions, I enhanced learnability, usability and effectiveness in task completion.
‍

Improved user journey through task simplification.

Applied data visualization techniques to identify LLM failures promptly.

Enabled users to navigate easily between thousands of conversations.

Facilitated rapid creation of monthly reports, streamlining process for increased efficiency.

Reused established UI components to enhance user learnability and streamline development processes.

Introduced new components for specific needs that existing components could not fulfill for effective user experience.

Project challenges

Facilitation of user engagements. Engaging users effectively was challenging. Coordinating schedules and gathering meaningful feedback required careful planning. I learned to facilitate an comfortable environment for everyone for honest input.

Designing with LLM. Balancing technical constraints of LLMs with user expectations was demanding. Through iterative testing and close collaboration with, I created an efficient user experience.

Adapt to new design system. As a new member of the team, it was challenging to quickly adapt to the established design system. By thoroughly studying the guidelines and iterating designs based on feedback, I managed to ensure consistency while bringing in innovative solutions.

Key takeaways

Strategy. Every task should have a clear purpose and understanding of its future evolution, enabling effective strategy and design planning that aligns with the company’s goals.

Proactivity. Self-initiative and continuous curiosity are crucial for problem-solving and knowledge acquisition. “The only thing stopping me from doing a better job is myself.”

Forgiveness over permission. Voice my expert opinion and take necessary actions, even if it requires post-approval while respecting constraints. “Don’t ask for permission but for forgiveness. I am an expert and need to voice my opinion.”

Inclusive stakeholders. Engage stakeholders early to incorporate diverse perspectives and enhance product development. “A fantastic way to start a conversation with someone is by asking them what they think of your idea and why they are interested in it.”