menu-squared-2

Overview: Redesigning Google's Customer Support Data Catalog

I was hired to redesign Google's Customer Support Data Catalog, with the goal of integrating advanced AI and machine learning technology to streamline the handling of vast customer support data for all Google products, worldwide. The project planned for short and long term improvements by enhancing usability and data discoverability through user-centered design methods and new advanced generative AI vector database and data lake technology.

Problem: UX Audit Shows Poor Usability And Data Access

The core challenge lay in the catalog's convoluted usability stemming from a lack of consistent onboarding and uniform documentation across its vast, legacy data repositories. Users were frustrated with an inefficient search experience, compounded by legacy architecture that complicated data discoverability and accessibility. This led to a frustrating user experience for those engaged in customer support reporting, BI analytics and machine learning model development, necessitating a redesign that not only streamlines search functionalities and data organization with generative AI, but also seamlessly integrates machine intelligence into the user flows to alleviate technical and navigational pain points universally.

My Role

UX Researcher
Lead UX Designer

Programmatic Infrastructure Engineering team

Tools

Figma, FigJam, Miro, Adobe Illustrator, Photoshop, Google Workplace Tools

UX Research Discovery Files in Figjam: UX Research Plan, UX Interview Script, Scheduled User interviews, Individual user interviews with analysis, Customer Journey maps and flows.

User Research And Strategic Planning

By adopting a holistic research approach, my work emphasized strategic planning, in-depth user research, and a profound research engagement with the data catalog's impact on users' lives. This expansive view of UX research fostered a comprehensive understanding of the legacy data catalog ecosystem, enabling the development of AI generative retrieval solutions that were deeply rooted in user’s data needs, activities and the role that they played in their particular team and their role in the company.

User Interviews: User Pain/Frustrations

Employing the "critical incident" method, I conducted user interviews to capture vivid accounts of challenges and triumphs faced by users. I conducted 18 user interviews over a 3 week period with a mix of novice to advanced data catalog users and uncovered three main persona users. I documented the key user frustrations, pain points and insights the users had using the data catalog. These conversations offered a comprehensive view of user issues, allowing me to challenge and test existing assumptions.

User Experience Problem Areas
  • Difficulty searching data sources and filtering on results
  • Challenges in data discovery and data comprehension
  • Obstacles with locating and accessing data sources, lack of consistent business logic, data definitions and metadata
  • Inconsistent data onboarding – naming field titles, metric definitions, signals
  • Time consuming and difficult to access data and metadata
  • Lack of visibility into data freshness and quality make it hard to trust data and low confidence in data single source of truth

I wrote a UX research plan and user research script to align with stakeholders on the key research goals. To be transparent, I included a full transcript of my interviews, analysis of user observations, user frustrations and actionable insights that I gained from users. The consistent way that I documented the user interviews helped me to identify the prevailing patterns and user trends which helped focus my personas and user Journeys.The output of this research became the application requirements for my UX design work.

Synthesize & Analyze: Identify Patterns

Following the user interview phase, I engaged in synthesis and analysis to categorize observations and identify prevailing user patterns and user trends. I used AI to help with my analysis, which was crucial in saving time and resources in translating my user research into actionable insights, giving me more time to focus on design strategy.

Affinity Diagram: Organize Ideas And Data

Collaborative sessions involved using bright colored stickies to capture the key pain points and insights, encouraging creativity and innovation as persona user flows are mapped and optimized.

Discover, Understand, Analyze And Learn

DISCOVER Phase: Simplifying Search and Browse: Users faced difficulties in searching for data and navigating the catalog. Focus was on creating an intuitive search and browsing experience.

UNDERSTAND Phase: Enhancing Comprehension and Data Interaction: Users desired easier data preview and documentation access, highlighting the importance of trust in data sources for effective usage.

ACCESS Phase: Streamlining Data Management and Addressing Access Complexities: Challenges in data entity access and tool management were noted, emphasizing the need for a streamlined access and management process.

LEARN Phase: Empowering Users and Enriching Data Consumption: User feedback revealed a need for better data understanding and the potential interest in a voice activated data chatbot, where users could actually talk to data.

As I continued my design thinking process of affinity mapping the user pain points, frustrations and user insights, the category clusters helped to map out each of the persona user journeys.

Personas And User Journey Flows

User feedback informed the development of three personas. I found critical user issues and challenges in data management, how data gets updated and the onboarding new data - all crucial issues that needed a solution.

As my customer journey mapping process was transparent and collaborative, the entire team gained a better understanding of the primary user experiences, ensuring that design decisions were consistently aligned with user needs.

Vision of a Future AI Experience

The hackathon (below) proved that a voice activated generative AI retrieval-based framework could be engineered for the data catalog - Transforming the “why” questions about your data into actionable “what” questions. The idea was to move away from the current manual data catalog archival system, and move into a generative retrieval AI role-based system that understands user context.

First place Winner in Google’s 24 hour AI Hackathon

I was the sole designer working with five Google engineers, winning the "AI-Pivotyness" category. Our small team designed and built a working prototype of a voice activated data bot - a generative AI retrieval search engine for the customer support data catalog in two very long development days.

(Google Programmatic Infrastructure Engineering Hackathon May, 2023)

As the user persona journeys converged and overlapped, the research showed that we needed to focus on the novice users of the data catalog.

Stakeholder Presentation: Aligning Vision

The presentation of my UX research findings was pivotal in aligning stakeholder project objectives with my user research. With the buy-in from the key stakeholders, it ensured that data catalog design decisions were grounded in user research, enhancing the user experience across the various data catalog touch points

I categorized observations to identify trends and insights into a research presentation for stakeholders, which served to align application goals and act as a foundation for design strategy.

Ideate Future Vision: Refine Concepts

Transitioning from ideation to implementation, I developed low-fidelity prototypes from my sketches and concept drawings to explore diverse design solutions. This iterative process involved engaging with cross-functional team members and to refine the concepts based on feedback, ensuring that the designs were simple, easy to use and addressed the user needs and challenges that I identified in my research.

Implement Ideas: Rapid Prototyping

Throughout the project, a focus on maintaining a consistent, high-quality design process that was adaptable to various user and internal business demands ensured the delivery of a user-centered, redesign of the data catalog.

Customer Outcome: Innovative Solution

This detailed case study encapsulates the comprehensive user journey undertaken to redesign the Customer Support Data Catalog, highlighting the intricate challenges, innovative solutions, and the strategic, user-focused design process that culminated in an enhancement of better usability and data discoverability through user-centered design methods and new advanced generative AI capabilities.

Evaluate & Improve: Continuous Feedback

A beta testing phase, coupled with continuous user feedback and performance analysis using Google employee feedback, facilitated iterative improvements, continually refining the Customer support data catalog

Next case study Back to homepage