A sales leader opens the CRM to answer a simple question. Are we on track this quarter?
They filter pipeline by region, then by stage. Something looks off, so they drill into a few accounts. Each click is another question. Each click should be instant.
But it’s not. Under load, latency climbs, and spinners appear. The experience breaks and trust drops.
It’s a concurrency problem on live data. When you’re embedding analytics, hundreds or thousands of users explore the same tables at once. If the response is slow, your product feels slow.
Traditionally, teams build complex workarounds to try to manage this. Tables and schemas are tuned to eke out performance gains, data gets copied, views are precomputed, and extra caching layers are added. These approaches work, but they add operational overhead and make systems harder to scale, especially as data volumes and use cases expand.
That’s why when we dug into Snowflake’s Interactive Analytics, we were excited about the benefit for our joint customers. Interactive Analytics, powered by interactive tables and warehouses, supports real-time exploration and keeps queries fast and predictable. For product and data teams building analytics experiences, that means filters update instantly, drill-downs feel natural, and performance holds up during peak usage like end-of-quarter reviews. You no longer have to do a lot of tuning and architecture gymnastics to hit SLAs 😅.
When paired with Omni’s AI analytics platform, this setup makes the experience even stronger. Snowflake’s Interactive Analytics handles the heavy lifting on query execution, while Omni focuses on how users actually interact with data. Revenue, pipeline, and quota metrics are defined once in Omni’s semantic layer and pushed back to Snowflake, or pulled in through our integration with Snowflake semantic views. Those same definitions power natural language chat, dashboards, spreadsheets, and SQL queries.
In the demo below, I walk through this in practice. First, I compare performance using Omni’s dashboard performance profiler, running the same revenue and pipeline queries against different table setups. Then I show how those reports are embedded into a CRM inspired application called Dealio. The result is a branded analytics experience where thousands of sales leaders can filter pipeline, drill into deals, and ask follow-up questions without being slowed down.
Sales leaders can explore data freely, confident that the numbers they’re seeing are consistent and governed by the right business logic and permissions. Product managers don’t have to worry about concurrency, and backend engineers don’t have to build and maintain custom optimization layers just to keep things fast.
This shift isn’t just about shaving milliseconds off a query. It changes how people use analytics. When responses are instant, analytics becomes part of the decision-making workflow instead of something you wait on. It means being able to serve thousands of stakeholders without the performance and complexity tradeoffs that used to make those experiences feel out of reach.
For a deeper look into how you can build & deploy your own AI embedded analytics product using Omni, check out this recent webinar with our product team. They walk through how we optimize data for AI, built a branded experience, and ensured security and performance.