Unify admin’s tool ecosystem

Role

UX lead, research, vision development

Timeline

2 years

How might we build an intelligent, adaptable & holistic admin surface that optimizes for users’ goals and guides them through complex journeys?

A problematic ecosystem

  • Complex scale: 30M+ customers, 10k+ agents, 250+ products from 40+ organizations across Google

  • Fragile & inefficient: fragmented and error-prone tools hinder agility and slow response to changing business needs

  • Low admin satisfaction: Low CSAT scores in quarterly tool surveys

Diving into the unknown

This project was the inaugural effort in unifying the experiences across all the support admin roles. I partnered with research to survey the admins about their primary goals and tasks. A followup survey on a list of specific journeys helped to pinpoint the most problematic admin workflows.

Together with a group of support admins, I ran multiple journey mapping sessions to help the team visually identify opportunities for the different stages of the admin workflows.

Sketching the new experience

I ran multiple cross-functional workshops with business and admin partners to align on the problems to tackle. We used activities like empathy mapping and creative matrix to brainstorm opportunities and solutions.

I then led multiple iterations of design, starting with low-fidelity mocks to get buy-in from our internal partners. This was followed by redesigning the IA to support complex cross-tool journeys. Finally, we validated the wireframe concepts with the admin users.

Socializing the story

I created a vision story to weave together the concepts and highlight the admin pain points the design solved for. This story was socialized broadly within Google Support and partner teams running support operations for Ads, YouTube, etc.

Executing the vision

Key elements of the vision–system health, ecosystem hygiene, and configuration accuracy–directly shaped our roadmap. Subsequent feature launches addressed these problems, aligning the product with important business objectives and brought the product closer to the vision.

We also began an ML exploration to tackle critical routing problems which resulted in poor system hygiene. I partnered with PM and tech lead to generate and test ML concepts with admins, using the feedback to inform an iterative ML experimentation process. This work culminated in us launching predictive routing to certain business verticals for the MVP, covering about 30% of Google Ads issues. Early data showed ~2% increase in routing accuracy and a 20 min reduction in median resolution time.

Happier support admins!

Over a period of 2 years, we improved the admin experience, resulting in a 10% increase in CSAT. We also made measurable business impact by enabling admins to accurately configure support in response to changing business needs.