CASE STUDY: 2023 – 2026
Accelerating Cloud Assist Adoption: Guardrails and Governance for Enterprise GenAI
Establishing secure admin defaults and policy architectures to unblock enterprise cloud AI adoption
Gemini Cloud Assist: Enterprise AI Governance, Administrative Controls & Unified SKU Alignment
Key Metric
110K+ MAU
Exceeded the 100K target of monthly active users.
Retention
>60% Retention
Achieved a >60% 90-day retention rate for Enterprise users.
Satisfaction
>50% CSAT
Based on in-product surveys for active users.
Efficiency
Nearly 2× faster RCA
Nearly doubled the speed of root cause analysis for customers troubleshooting cloud issues.
Unifying fragmented AI utilities into a cohesive SKU while unblocking launch velocity and establishing scalable administrative governance.
Supported critical strategic goals as the UX lead to ensure the platform’s business and engineering viability. Analyzed technical intricacies across Workspace and GCP resource models to establish foundational administrative frameworks.
Executive Summary
“Gemini Cloud Assist has significantly helped our dev teams. It reduced the number of outreach and touch points I have with them regarding GCP questions by 60%. This allows our cloud team to scale more effectively and focus on more complex tasks.” — Customer
The challenge required overcoming engineering bottlenecks for the Chat UI team, establishing organizational-level API and IAM governance, and unifying standalone AI utilities into a single, cohesive SKU without reinventing existing security infrastructure.
Project Goals & Strategic Focus
Unblocking GA Launch Velocity for the Chat UI Team
Authored foundational UX specifications to coordinate cross-functional development, without which the engineering team explicitly noted the launch of Gemini Cloud Assist would have been delayed.
- check_circle Spec Authoring & Coordination Docs
- check_circle Cross-Functional Walkthroughs
Unifying Standalone “Add-Ons” Into a Single Gemini SKU
Conceptualized and facilitated a major cross-product alignment workshop bringing together UX, PM, and engineering leads from Databases, BigQuery, Gemini Code, and the Gemini Design System to fold fragmented legacy AI utilities into one cohesive SKU.
- check_circle 4+ Product Area Workshops
- check_circle End-to-End Lifecycle Experience
Preventing Duplicative Engineering & Challenging Assumptions
Advocated basing the Gemini enablement and settings interface on existing standards, saving thousands of engineering hours, and defined permission models that align with policy and security standards.
- check_circle Standard Security Framework Reuse
- check_circle Best-Practices Spec for ENG
Deconstructing enterprise administrative gatekeeping and end-user data transparency requirements.
Research and discovery revealed that administrators act as critical gatekeepers for enterprise readiness, compliance, and feature enablement, while end-users require clear agency and transparency over their personal data boundaries.
Enforcing Enterprise Grounding & Compliance
Admins need strict controls to dictate the type of grounding used. For highly regulated environments, they must be able to enforce Enterprise-friendly grounding over standard web grounding based on global search.
Release Channels & Pre-GA Feature Access
Admins require the ability to toggle between General Availability and Preview release channels, allowing them to safely test Preview features in non-production projects before rolling them out widely.
Page Context Preferences & Data Agency
End-users must have explicit agency over their personal data boundaries, including opting in or out of including current page context as a chat data source via a dedicated Gemini user preferences menu.
Actionable Transparency on Blocking & Data States
When end-users lack access, the UI must explain whether it is due to API enablement status or lacking specific IAM permissions needed to enable APIs rather than generic errors.
Harmonizing administrative configuration, grounding guardrails, and quota transparency.
Structuring the end-to-end user experience required balancing administrative controls with end-user onboarding and transparency.
Cloud Assist Enablement & Settings
Ensuring Seamless Adoption of Cloud Assist’s Full Capabilities
Ensuring users have privacy by default but are also encouraged to adopt new features while sharing their usage data to help Google improve the UX longer term.
Enabling Cloud Assist onboarding from Compute Engine
Governing Release Channels and Web Grounding
Allowing administrators to easily switch existing preview customers to the Generally Available version, or conversely, grant teams access to Preview features in non-production environments. Admins can also enforce strict enterprise web grounding for highly regulated projects.
Admin settings: grounding, release channel & data sharing
Actionable Blocking Explanations and Data States
Giving clear error messages to users who lack access. This includes explaining when they are blocked from using features due to missing API enablement status or lack of admin permissions to enable APIs.
API enablement status before setup
Communicating Missing Permissions
Ensuring users clearly understand when they are missing permissions to use Gemini Cloud Assist and understand what to do next.
Panel explaining missing permissions that block Cloud Assist enablement for non-admins
Exceeded Monthly Active Users Target and Drove >60% Enterprise Retention
The product rapidly gained adoption across developer and engineering teams, surpassing the ambitious 100K Monthly Active User goal while establishing exceptional long-term retention among enterprise customers.
Results
- >110K Total MAU: Exceeded the goal of 100,000 Monthly Active Users.
- Enterprise Growth: Exceeded benchmarks for Enterprise active users and non-Enterprise users.
- >60% Enterprise Retention: Surpassed the 60% retention target for 90-day retention rate for the Enterprise users.
Optimized CSAT and Support Case Deflection
The product’s performance and business impact were measured heavily by user satisfaction and support deflection, providing AI-driven self-service troubleshooting that dramatically streamlined developer workflows.
Results
- >50% CSAT: Achieved a CSAT score above 50%, tracking toward the overarching >60% satisfaction goal.
- >30% Support Deflection Goal: Established AI-driven self-service troubleshooting capabilities designed to deflect customer support cases.
What product, research, and design leaders said about the work.
“Congratulations on the successful launch of Gemini Cloud Assist at Next ’24!!! It would not have happened without all your hard work and focus and we thank you for all your efforts.”
VP & GM, Google Cloud Developer
“Thanks for your various hard work ensuring the successful launch of Code Assist free trial, Admin Control, and various Cloud Assist launches. Your work here has been critical to Admin for Gemini’s success, not only cranking out updated designs at a breathtaking pace but also on hand to answer questions from the front-end team all hours of the day, and your attention to detail meant that no stones were left unturned in our never ending quest to ensure a quality product.”
Senior Product Manager, Admin for Gemini
“I want to extend my gratitude to Lane for her tireless leadership in working through Gemini and GCA Enablement, particularly in all the work she did, and is still doing, with the outcomes from the workshop. Without her it wouldn’t be a success, she is the leadership glue that’s holding all the disparate teams and ideas together and driving the vision forward.”
Senior Staff UX Manager, Google Cloud Databases
“Thanks to Lane for jumping in to set up Cloud Assist test accounts and troubleshoot all of the permissions issues. Unexpected issues arose while I was in the middle of a diary study and getting frantic emails from participants not able to access their accounts. Lane’s fast support and technical expertise helped save the day and made the rest of the study go smoothly.”
Senior UX Researcher, Cloud Assist
Empowering Scalable, Secure Enterprise AI Access
By establishing foundational administrative controls, unifying fragmented AI add-ons into a cohesive SKU, and reusing proven security frameworks, Gemini Cloud Assist empowered 110K+ monthly active users while setting a new benchmark for cross-functional launch velocity and enterprise trust.