Okay, I’ve rewritten the content to sound more human and natural, while carefully preserving every single HTML tag and its attributes exactly as they were provided.
AgentStack is here to tackle the biggest challenge in enterprise AI: getting multi-agent systems up and running smoothly without tying developers down to just one cloud or data platform.
</figure>
</div>
</div>
</div>
</div>Teradata has taken its agent-building tools, first introduced last year, and turned them into a comprehensive toolkit. They believe this will really help businesses overcome the hurdle of moving AI agents from experimental projects to fully operational, production-ready systems.
Now called Enterprise AgentStack, this upgraded toolkit adds AgentEngine and AgentOps to Teradata’s original Agent Builder. That builder already came with a user-friendly interface for creating agents, using handy third-party tools like LangGraph, and some smart context intelligence features.
Think of AgentEngine as the power behind deploying agents across various IT setups, whether on-premises or in the cloud. AgentOps, on the other hand, gives you one central place to find, keep an eye on, and manage the entire lifespan of your agents throughout the company.
According to Stephanie Walter, HyperFRAME Research’s AI stack expert, the AgentEngine is absolutely essential to Enterprise AgentStack. It’s the crucial link that bridges the gap between designing an agent and actually making it work in the real world.
Walter explained, “Without an execution engine, companies usually have to piece together custom code just to get their agents to talk to each other. The Agent Engine streamlines how agents run and helps businesses truly grasp their performance, reliability, and potential risks, even at a large scale.” She emphasized that this kind of AgentEngine capability is exactly what companies need to successfully get agents or agentic systems into live production.
But here’s where it gets interesting: analysts point out that Teradata’s strategy for helping businesses adopt agents is quite different from what competitors like Databricks and Snowflake are doing.
Robert Kramer, a principal analyst at Moor Insights and Strategy, explained the differences: “Snowflake has been pushing its Cortex and Native App Framework, allowing companies to develop AI apps and agents right next to their carefully managed data. Databricks, on the other hand, is all about agent workflows within Mosaic AI, focusing heavily on building, orchestrating, and evaluating models, all connected to their lakehouse system.”
Echoing Kramer’s thoughts, Walter highlighted Teradata’s unique selling point: Enterprise AgentStack is designed to be a flexible execution and operations layer that works with *any* vendor and across mixed IT environments. This means it doesn’t force agents to be tied to just one cloud or data platform.
Walter further elaborated that this flexibility comes from Teradata’s use of popular third-party tools like Karini.ai, Flowise, CrewAI, and LangGraph. This approach allows businesses and their developers to adapt and grow their agent systems without getting trapped in proprietary platforms from companies like Snowflake and Databricks, which often aim for total control within their own ecosystems.
However, Walter also offered a word of caution: while Enterprise AgentStack’s design is a great fit for businesses, its real test will be whether it can consistently maintain strong integrations with all those third-party frameworks.
She stressed, “Customers will definitely want solid proof that AgentStack can handle complex, continuous, multi-agent deployments in a live production setting.”
Kramer also advised that businesses and developers should really dig into how easy and practical AgentStack is to use *before* committing to it.
He elaborated, “It’s crucial to see how smoothly you can apply consistent rules, test things after updates, track down issues from start to finish, and connect it with your current security and compliance systems. Being ‘open’ is great, but only if it doesn’t just push more complexity onto the customer.”
You can expect Enterprise AgentStack to enter a private preview phase for both cloud and on-premises use sometime between April and June of this year.
