Skip to main content
Eye on AI

#327 Baris Gultekin: The Next Phase of AI - Agents That Understand Your Company's Data

42 min episode · 2 min read
·
Baris Gultekin

Episode

42 min

Read time

2 min

Topics

Marketing, Sales & Revenue, Artificial Intelligence

AI-Generated Summary

Key Takeaways

  • AI Governance Architecture: Run AI models inside the Snowflake security boundary rather than sending data to external APIs. Snowflake hosts models from OpenAI, Anthropic, Gemini, and Meta within customer cloud environments on AWS, Azure, or Google Cloud, ensuring no data is stored by model providers or used for training, and existing data access controls automatically apply to all AI outputs.
  • Structured Data Retrieval as Differentiator: Text-to-SQL generation for structured data is significantly harder than unstructured document retrieval. Enterprises should invest in semantic models that capture business-specific data definitions before deploying agents. Without accurate, maintained semantic models, agents produce unreliable answers to factual business questions like monthly revenue figures, where only one correct answer exists.
  • Agent Deployment Maturity Model: Production agent rollouts follow a four-stage sequence: proof of concept, small pilot, broad deployment, and continuous optimization via feedback loops. Enterprises currently operate hundreds of agents at most, not thousands. Agent memory capabilities now allow systems to learn from usage patterns and self-correct over time, reducing manual developer intervention between deployment stages.
  • Data Preparation as AI Prerequisite: Before building any agent, enterprises must consolidate data from siloed sources, assign semantic definitions, build search indices, and process unstructured content into structured formats. Snowflake calls this making data "AI ready." Organizations skipping this step encounter retrieval failures regardless of model quality, since AI output quality is bounded entirely by the quality of accessible data.
  • Democratization Replacing the Middle Layer: The translator role between business expertise and technical data systems is disappearing. Natural language interfaces now allow non-technical employees across sales, marketing, finance, and the C-suite to query governed data directly. One Snowflake customer eliminated 2,000 hours of manual call-center analysis work by deploying a single data agent against existing call records.

What It Covers

Baris Gultekin, Snowflake's Head of Product for AI, explains how Snowflake builds enterprise AI agents that operate directly within governed data environments, covering the architecture behind Snowflake Intelligence, structured data retrieval challenges, agent reliability frameworks, and why data preparation is now the prerequisite for any viable enterprise AI strategy.

Key Questions Answered

  • AI Governance Architecture: Run AI models inside the Snowflake security boundary rather than sending data to external APIs. Snowflake hosts models from OpenAI, Anthropic, Gemini, and Meta within customer cloud environments on AWS, Azure, or Google Cloud, ensuring no data is stored by model providers or used for training, and existing data access controls automatically apply to all AI outputs.
  • Structured Data Retrieval as Differentiator: Text-to-SQL generation for structured data is significantly harder than unstructured document retrieval. Enterprises should invest in semantic models that capture business-specific data definitions before deploying agents. Without accurate, maintained semantic models, agents produce unreliable answers to factual business questions like monthly revenue figures, where only one correct answer exists.
  • Agent Deployment Maturity Model: Production agent rollouts follow a four-stage sequence: proof of concept, small pilot, broad deployment, and continuous optimization via feedback loops. Enterprises currently operate hundreds of agents at most, not thousands. Agent memory capabilities now allow systems to learn from usage patterns and self-correct over time, reducing manual developer intervention between deployment stages.
  • Data Preparation as AI Prerequisite: Before building any agent, enterprises must consolidate data from siloed sources, assign semantic definitions, build search indices, and process unstructured content into structured formats. Snowflake calls this making data "AI ready." Organizations skipping this step encounter retrieval failures regardless of model quality, since AI output quality is bounded entirely by the quality of accessible data.
  • Democratization Replacing the Middle Layer: The translator role between business expertise and technical data systems is disappearing. Natural language interfaces now allow non-technical employees across sales, marketing, finance, and the C-suite to query governed data directly. One Snowflake customer eliminated 2,000 hours of manual call-center analysis work by deploying a single data agent against existing call records.

Notable Moment

Gultekin describes using Snowflake's own agent internally as a product manager, replacing a multi-day data scientist analysis cycle with a seconds-long natural language query. He frames this not as automation but as a cultural shift in how organizations operate when data access becomes universal.

Know someone who'd find this useful?

Episode Transcript

Can you walk me through what it's been like? I mean, the last two or three years have just been crazy, and it doesn't seem to be letting up, and how Snowflake AI has evolved. AI is, incredibly exciting. So we've been building a lot of product over the last two, two and a half years. Anywhere from building AI to run right next to data to just being able to kind of talk to your data in natural language and and democratize that access and and glean a lot more insight from data. So that's the high level goal of the set of products that we're building, and that's something. Most AI is just speech to text plus a language model. It's full for reading transcripts, not understanding conversations. Velma from Modulate, an AI built on ensemble listening model architecture, specializes in audio analysis. It orchestrates hundreds of smaller sub models purpose built to understand the nuances of voice like tone, timing, stress, and intent. Perfect for fraud defense, deep fake detection, agent attrition prevention, or customer service moderation. Check out the live Velma preview at preview.modulate.ai. That's preview.modulate.ai to see how the model breaks down audio providing time stamped explainable signals. Stop transcribing. Start listening with modulate.ai. Hi, Craig. Nice to meet you. Thanks for having me here. So my name is Barish. I'm the head of product for AI at Snowflake. I joined Snowflake about two and a half years ago when Snowflake acquired my company. And before that, I was at Google for a long time, most recently running AI initiatives focused on the Google Assistant Google Assistant products for for multiple years, for a decade. Before, AI was super hot. I love I love the space. I've been in the space for a long time. And at Snowflake, my my job is to build out Snowflake's AI strategy as well as a series of products that we're building for our customers. AI is incredibly exciting. So we've been building a lot of product over the last two, two and a half years. You know, anywhere from building AI to run right next to data to just being able to kind of talk to your data in natural language and democratize that access and glean a lot more insight from data. So that's the high level goal of the set of products that we're building at Snowflake. Yeah. And I have to apologize. I said Boris. I'm sure you get that along a lot. It's pronounced Barish? It's Barish. Yeah. Barish. Okay. And can you walk me through what it's been like? I mean, the last two or three years have just been crazy, and it doesn't seem to be letting up. And how Snowflake AI has evolved? Yeah. Certainly. I mean, the journey started about two and a half years ago. With the feedback we got from our customers is AI is incredibly transformative, but, of course, governance is also equally important. And for Snowflake, …

Get the full transcript (6,859 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all Eye on AI transcripts →

You just read a 3-minute summary of a 39-minute episode.

Get Eye on AI summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links.

Tools

  • by Snowflake

    Baris Gultekin, Snowflake's Head of Product for AI, explains how Snowflake builds enterprise AI agents that operate directly within governed data environments, covering the architecture behind Snowflake Intelligence
  • by Modulate

    SPONSORS: Modulate (Velma)

More from Eye on AI

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best AI Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's AI & Machine Learning Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into Eye on AI.

Every Monday, we deliver AI summaries of the latest episodes from Eye on AI and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime