How Fireblocks Uses Snowflake AI Agents to Handle 40-50% of Data Queries
Fireblocks is the digital asset infrastructure provider powering $10T in transactions across 550M crypto wallets. The company deployed Snowflake’s AI Data Cloud with Cortex Agents to unify 15 data domains and automate analytics for both customers and internal teams. The result: AI agents now handle 40–50% of all data queries, saving the equivalent of two full-time analysts per month.
Impact
40–50%
Share of data queries handled by AI agent
2 full-time equivalents
Analyst capacity saved per month
Challenge
With billions of rows of data spread across 15 separate domains, basic analytics queries were taking up to 20 minutes and the data team could not keep up with internal demand, while customers lacked direct, secure visibility into their own transaction data.
Solution
Fireblocks unified all data on Snowflake’s AI Data Cloud running on AWS, then built AI agents using Snowflake Cortex Agents with semantic views—both an external customer-facing agent (Fire Genie) and an internal analytics agent—enabling natural language queries across all data domains with role-based access controls enforcing data isolation.
Tools & Technologies
AWSWhat Leaders Say
“Snowflake wasn’t a ‘nice to have’ for us. It was an absolute must so we could continue growing.”
“Once Snowflake gave us access to Cortex, our world got a whole lot easier.”
“With our internal agent and Snowflake, teams can go from a Jira ticket to the deepest analytics on product metrics in one place — all with correct semantic views.”
Sign up to read complete case studies, access detailed metrics, and unlock all use cases.
Full Story
Fireblocks sits beneath a significant portion of the world’s digital asset economy, providing the infrastructure that secures cryptocurrency transactions and stablecoin flows for hundreds of millions of end users. At the scale the company operates—billions of rows of on-chain and transactional data spread across 15 distinct data domains—keeping analytics responsive and accessible was becoming an existential challenge rather than a nice-to-have.
Before adopting Snowflake, Fireblocks was hitting hard limits. Basic queries against tables with billions of rows were taking up to 20 minutes to return. The data team was fielding a volume of requests from product, sales, and engineering teams that manual analysis could no longer satisfy. Siloed data domains meant a sales representative seeking revenue insights and a product developer needing staking account data had to go through entirely different pipelines—with no shared context between them.
The company rebuilt its data platform on Snowflake, running on AWS, and began routing all data—from Salesforce CRM analytics to Zendesk support insights—through the AI Data Cloud. With Snowflake’s unified semantic layer across all 15 domains, teams across the business gained a single, queryable source of truth. The most significant step forward came in a 45-day internal hackathon: engineers built Fire Genie, an AI agent powered by Snowflake Cortex Agents with semantic views, which lets customers query their own wallet and transaction data in natural language through a secure API. A parallel internal agent, built on the same architecture, extended the same capability to Fireblocks’ own staff. Streamlit dashboards on AWS allowed teams to share live, auto-updating reports in place of static screenshots.
The numbers from the internal agent alone tell a compelling story. It now handles over 2,000 analysis queries per month, accounting for 40–50% of all analytical queries company-wide—saving an estimated two full-time analysts’ worth of capacity every month. The data team shifted from being a bottleneck for basic requests to focusing on advanced insights that directly influence product development and customer retention.
Fireblocks is now expanding Fire Genie to let customers build personalized dashboards, while pushing toward a single internal self-service analytics environment where any employee can take a query from initial idea to complete final report without waiting on the data team. For a company whose competitive advantage depends on helping clients trust and understand their digital assets, the ability to put accurate, real-time data in anyone’s hands is not just operational improvement—it is product differentiation.