Real-Time Context Is Not Optional: Why Batch Data Breaks AI Agents
Batch data breaks AI agents in production. Real-time context ensures fresh, reliable decisions powered by CDC, streaming, and data products.
Hosted by Saket Saurabh, CEO of Nexla, each episode features conversations with CDOs, CTOs, and data architects who’ve built systems at scale. No vendor pitches. No theoretical frameworks that fall apart in production. Just honest discussions about what works when you’re dealing with data variety, stuck AI pilots, or architectures that somehow got more complex in the cloud.
Most data content today falls into two camps: polished conference panels that skip the hard parts, or vendor-funded thought leadership that’s really just a sales pitch in disguise.
This podcast exists in the white space between them.
Every episode breaks down real architecture decisions, team structures, and hard-won lessons from leaders who’ve actually shipped. For example, why most AI adoption stalls at the process level, how to handle data variety across 1000+ business systems, or when to use fine-tuning vs RAG for production agents — the stuff that doesn’t make it into the case studies.
You’ll hear about:
New episodes drop every two weeks on Apple Podcasts, Spotify, and YouTube.
Episode 1 features Robin Sutara, Field Chief Data Strategy Officer at Databricks, on why AI adoption stalls and how to fix it at the process level.
Coming up this season:
Subscribe wherever you listen to podcasts, or watch full episodes on our YouTube channel.
If you’re a data leader with a story worth sharing, reach out. We’re always looking for guests looking to share their lessons learned.
Batch data breaks AI agents in production. Real-time context ensures fresh, reliable decisions powered by CDC, streaming, and data products.
Bigger context windows do not always improve AI agents. Learn why targeted context engineering delivers better enterprise AI performance.
Learn how to automate F5Bot alerts into Slack using Nexla to track Reddit and Hacker News mentions in real time.