Agentic Data Cloud Brings System of Action to Life
The way Google handles the underlying data layer is completely changing. Instead of acting as a giant digital storage locker where information just sits there, your data environment is now an active engine.
Google refers to this as a “System of Action,” meaning it’s specifically built for AI agents to constantly read, update, and act on your information in real time.
Yes, this is the fourth part of our special blog series. Welcome back!
This AI-native vertical integration is what allows enterprises to scale from human-led analysis to agent-driven execution without the typical fragmented stack costs or hallucinations.
For businesses, this means less time spent moving and preparing data, fewer fragmented systems, lower operational costs, and much faster decision-making across teams and clouds alike.”
Goodbye Dataplex, Hello Knowledge Catalog
The core of this new setup is the Knowledge Catalog, which officially replaces Dataplex. It pulls in information from all over the place, including outside platforms like SAP, ServiceNow, and Palantir, and actually figures out the business context behind it.
A dedicated LookML Agent automatically handles the technical definitions, and a massive hybrid search system makes sure you can find exactly what you need without breaking your existing security rules.
On top of that, a feature called Smart Storage kicks in the second an unstructured file, like a random PDF or image, drops into your Cloud Storage. It instantly labels and organizes the file, so your team never has to do it by hand.
Ending the Cross-Cloud Headache
Google is also making it significantly easier to work across different clouds. With the Apache Iceberg REST Catalog standardization, your AI agents can now query information living in AWS, Azure, Databricks, or Snowflake without actually moving anything.
That means no messy transfers and zero extraction fees. Besides, Spanner Omni can now run across multiple environments, and AlloyDB connects directly to your lakehouse with no complicated pipelines to move the information around.
New Teammates
For developers and data engineers, the Data Agent Kit drops ready-to-use AI assistants directly into everyday workspaces like VS Code and the Gemini CLI. That said, you get specialized agents for specific jobs.
- Data Engineering Agent to handle pipeline changes.
- Data Science Agent to manage model lifecycles in Spark.
- Database Observability Agent that monitors your backend infrastructure 24/7.
Massive Speed & Cost Upgrades
To keep all of these new agents running smoothly, Google cranked up the performance across the board. The new Lightning Engine for Apache Spark processes data up to 4.5 times faster. Managed Lustre is hitting massive speeds of 10 TB per second, and Bigtable introduced a new in-memory tier with practically zero lag.
Best of all, BigQuery added a new fluid scaling feature that keeps your performance totally steady while actually dropping your costs by about 34%.
This means no egress fees and no more complex data pipelines just to make systems talk to each other. We’re finally moving toward a setup where data stays where it is, but is still fully usable.”
Ready to Take Your Data to the Agentic Level?
You can’t let your data just sit in a static storage locker anymore. If you want to go agentic and autonomous and capture real value, your infrastructure needs to become a live engine that constantly fuels your AI.
As a global Google Cloud Premier Partner, Cloudfresh helps you:
- Modernize data architectures for the agentic AI era.
- Migrate from Dataplex to the new Knowledge Catalog.
- Connect AWS, Azure, Snowflake, and on-prem data environments with BigQuery.
- Optimize BigQuery and Spark environments for performance and cost efficiency.
- Build AI-ready data foundations for analytics, automation, and autonomous agents.
Curious if your data is truly ready for the agentic-first era? Drop your details in the form below to schedule a quick call and assess exactly where your foundation stands today.
