Achieved scale of 500:1 user-to-employee ratio through self-service
10x the speed of analysis for end users through conversational analytics
Improved stability by migrating to Looker core on Google Cloud
Reduced manual reporting for 5,000 monthly active users
Google Cloud Support revolutionized operations, moving from fragmented tools to a governed, AI-powered Looker environment supercharged by conversational analytics.
Google Cloud Support revolutionized operations, moving from fragmented tools to a governed, AI-powered Looker environment supercharged by conversational analytics.
Overcoming metric drift with a
unified semantic layer
In a global enterprise, support operations inherently rely on data, and having immediate access to accurate statistics is vital for identifying case volume drivers and monitoring performance escalations. The Google Cloud Support BI team previously relied on a homegrown tool to manage this critical global support data. However, this decentralized approach mirrored challenges found in other legacy BI tools like Tableau, including significant metric drift, resulting in a lack of trust as different teams would define key performance indicators in conflicting ways. Without a governed set of measures, leadership lacked a single source of truth to monitor operational metrics or performance escalations effectively. This lack of governance made it nearly impossible to maintain a consistent 360-degree view of the customer experience across the organization’s nearly 5,000 monthly users.
Beyond data consistency, the team faced extreme scalability pressures. With a small team of full-time employees supporting thousands of Googlers and vendors, the BI team often became a bottleneck for data requests. The existing architecture lacked robust self-service capabilities, requiring a manual query for every new dashboard tile. To support a rapidly growing user base and highly complex products like BigQuery, the team needed a platform that could provide a stable, governed environment while empowering users to answer their own questions without constant technical intervention.
The team selected Looker for its semantic layer and centralized governance features. This transition allowed the team to move away from a fragile on-premise instance to Looker Core on Google Cloud, providing the stability required for high-concurrency workloads. By migrating to a unified platform, the team ensured that every role, from technical solutions managers to account executives, viewed identical numbers. This strategic shift was not just about changing software; it was about building a foundation for a data-driven culture that could operate at an unprecedented scale while maintaining absolute data integrity across all global support operations.
"We deal with thousands of incoming cases and are required to meet strict service levels to ensure customer production systems are up and running to meet their own needs. Conversational Analytics in Looker helps us get insights on how various shifts are run to ensure high operational rigor in solving customer cases," says Prathap Reddy, Managing Director of Google Cloud Support.
Looker’s semantic layer and governance is really important. We have leadership that needs everyone to have exactly the same visibility and understanding of what’s happening in the business—we can’t be looking at different numbers if we want to run our business or improve our customer experience.
Ron Farizon
BI Manager, Cloud Support Data and Analytics, Google Cloud Support
Empowering 5,000 users through
AI-driven self-service
With Looker as the foundation, the BI team successfully shifted to a "train the trainer" model, creating a self-service environment that dramatically increased operational efficiency. The impact was quantifiable as the ratio of users to BI staff jumped from 100-to-1 to a staggering 500-to-1. This efficiency is driven by Looker Explores, which allow non-technical users to build their own reports. For example, Technical Solutions Googlers now use the platform to monitor daily backlogs and performance metrics in real-time, while engineering teams leverage the data to measure product stability and identify bugs through a support lens.
The implementation of conversational analytics in Looker has further supercharged this self-service model. By connecting the conversational analytics agent directly to the existing, governed Looker data model, the team ensured that AI-generated insights remained fully accurate and compliant. The investment in metadata and data hygiene, originally intended for human users, became AI-friendly, paying off exponentially when the agent launched. Users can now ask complex questions in plain language, such as "Show me average resolution times for Gemini Enterprise Agent Platform cases," eliminating the need to hunt through dozens of bookmarked dashboards or manage complex filters.
Engagement has soared, particularly among senior leadership. High-level directors who previously found traditional dashboards too time-consuming are now top users of conversational analytics due to the flattened learning curve. This speed to insight has proven critical in high-stakes environments; for instance, leaders have been able to pull up live performance data during customer meetings to immediately address concerns. By moving beyond manual slicing, the team has turned data into a conversational asset that drives faster decision-making and more proactive support strategies for Google Cloud's massive global customer base.
"Conversational Analytics in Looker saves us 30 to 60 minutes daily, not only among leaders but across the managers community. Our insights have improved to target the most granular areas which are not working, helping us find the right solutions to issues quickly,” says Prathap Reddy, Managing Director of Google Cloud Support.
Conversational Analytics in Looker helps us see where hotspots are, which products are seeing a high incoming volume, monitor for backlog and escalations and finally review the customer experience scores to intervene and improve.
Prathap Reddy
Managing Director, Google Cloud Support
Future-proofing support with "AI on AI" insights
The roadmap for Google Cloud Support analytics focuses on expanding the semantic power of the conversational agent to tackle even more complex domains. While current implementations handle traditional BI metrics well, future iterations aim to integrate case text summarization and sentiment analysis. This would allow the agent to not only report on resolution times but also identify root causes and symptoms by "reading" through case histories. This semantic understanding represents a move beyond traditional BI, enabling the team to diagnose case spikes—such as those involving billing or token issues—with unprecedented speed and precision.
The team is also exploring an "AI on AI" use case by connecting the conversational analytics agent to upstream data, such as AI-powered chat concierge sessions. By analyzing these interactions, the support team can better understand the customer journey before a formal case is even created. Additional planned enhancements include automated anomaly detection and the integration of support personnel data, such as tenure and performance metrics. These tools will help managers proactively identify where additional training or resources are needed to maintain high service levels across global support sites and various vendor organizations.
"Because of Conversational Analytics and its ability to provide deeper insights quickly, our team has delivered 91% customer satisfaction for Premium support customers with an 88% AI-driven Sentiment score, which is run on 100% of cases, reduced escalation rates by 20% and achieved initial response time of 99.5% consistency over the last 6 to 8 months," says Prathap Reddy, Managing Director of Google Cloud Support.
By bringing in external context—such as standard operating procedures and support playbooks—the agent will eventually evaluate what percentage of cases followed proper protocols. Leveraging the governed Looker environment to feed these advanced AI workflows allows Google Cloud Support to set a new enterprise standard for agility. The ultimate goal is to shift from reactive monitoring to an automated ecosystem that shortens the gap between detecting an issue and taking corrective action.
I’m very excited for the ability to do repeatable analysis and anomaly detection. Having the 'what' and the 'why' immediately accelerates your ability to troubleshoot and take action. It turns a massive time-sink into a standalone, automated insight.
Ron Farizon
BI Manager, Cloud Support Data and Analytics, Google Cloud Support
The Google Cloud Support BI team manages data and analytics for global support operations, serving nearly 5,000 monthly active users across Google and its vendor partners.
Industry: Technology
Location: United States
Products: Looker, BigQuery, Gemini Enterprise Agent Platform
