❄️ Snowflake Cortex is Now Live in Astrato

Astrato now supports native integration with Snowflake Cortex LLM, giving you direct access to AI-powered insights — without moving a single row of data.

We’re proud to announce that Astrato is officially Snowflake Technology Ready Validated and a Snowflake Technology Partner. That means our integration meets Snowflake’s highest technical standards — and it’s ready for production use across your enterprise.

And yes — we’ll be at Snowflake Summit. Come see it live.

‍

⚡ What’s New?

With Cortex in Astrato, you get instant access to plug-and-play LLMs from leaders like Meta, Claude, DeepSeek, and Mistral. Prefer something custom? You can add your own model manually too.

But this isn’t just another chatbot UI bolted onto BI.

Astrato’s semantic layer and query engine work with Cortex to generate smart, context-aware SQL — not guesses. You don’t need to prep data, build pipelines, or handhold the model.

‍

🔒 Why Use Cortex in Astrato?

This is the first BI & Data App platform with in-warehouse LLM control and security.

• ✅ No data movement
  All your data stays in Snowflake — no syncs, no exports.
• ✅ Enterprise-ready
  Full RBAC via the CORTEX_USER role. Built for scale and trust.
• ✅ No-code required
  Just click to generate — models understand context via Astrato's semantic layer.

This is AI that fits your governance model, not the other way around.

‍

‍

👋 A Take from our resident Snowflake Data Superhero, Piers Batchelor

What excites me most about this Cortex integration? No data movement.
That one shift means Snowflake customers can start using LLMs instantly — no contract hurdles, no risky pipelines, no data leaving the warehouse. It’s not just fast; it’s enterprise-safe.

We’ve all seen the promise of “text-to-SQL.” And we’ve seen it fail.
The issue? Most tools don’t understand intent.
Ask for “sales of Apple products,” and you might get iPhones, or you might get Granny Smiths. Context matters. With Astrato + Cortex, you can now inject business context directly — or even bring your own fine-tuned model. That’s the difference between surface-level AI and something that actually understands your business.

Are LLMs perfect? No.
But they’re ready — if you treat them like a junior analyst. Tell them what you need, and why, and they’ll surprise you. Skip the context, and you'll get junk.

What Snowflake has done with Cortex is powerful. They’ve made fine-tuning easy. And that unlocks a new chapter in analytics — AI that lives in the data platform, and works for the business.

Now is the time.
We’re seeing more businesses build operational data products and interactive data apps — and AI’s role is growing fast. Think writeback validation, enrichment, categorisation — all done using LLMs, in real time, right where the data lives. 

‍

‍

‍

‍

‍

🤖 The Trouble With Text-to-SQL (And How We Solve It)

Text-to-SQL sounds simple. It’s not. Most LLMs choke on:

• Vague column names
• Schema sprawl
• Query context loss
• Over-fetching and bad joins

Astrato avoids those pitfalls entirely. We're doing it differently with AI powered by your semantic layer.

• ✅ Uses a curated semantic layer
• ✅ Injects context and metadata into the prompt
• ✅ Suggests, doesn’t guess — thanks to our engine
• ✅ No pipelines — just live SQL over Snowflake

You get trusted results, fast — with nothing lost in translation.

‍

‍

We’re excited to push boundaries with Snowflake Cortex. If you're heading to Summit, let’s connect. Otherwise, request a demo and we’ll show you what enterprise AI should actually feel like.

‍