Teradata launches AgentBuilder to boost secure AI deployment

Teradata has introduced AgentBuilder, a suite of tools aimed at supporting the development and management of autonomous, contextually intelligent AI agents within enterprise environments.

AgentBuilder is designed to help organisations address key challenges associated with deploying agentic AI, such as fragmented data, lack of integrated business knowledge, potential for unreliable outputs, excessive operational costs, and insufficient governance. These issues are particularly pressing for sectors such as financial services, where AI errors can result in considerable financial and reputational harm.

Overcoming AI deployment obstacles

AgentBuilder leverages open-source frameworks and operates on the Teradata AI and knowledge platform. It allows teams to design, operationalise, and manage multi-agent systems that can interact with Teradata Vantage's data resources, analytics, and hybrid infrastructure. The release includes ready-to-deploy Teradata Agents - templates that aim to speed up implementation and provide targeted solutions for specialised business challenges. These agents embed domain-specific logic into their workflows and are intended to produce more contextually relevant and reliable outputs.

According to Teradata, AgentBuilder will enable the transition from AI experimentation to full-scale, governed production deployments. By integrating contextual business knowledge and domain expertise, the suite aims to ensure that autonomous AI agents align with organisational goals and compliance requirements.

Key features and integration

AgentBuilder incorporates Teradata's Model Context Protocol (MCP) Server, which provides core infrastructure to support agentic capabilities. The MCP Server offers developers and AI professionals a curated set of prompts, components, and resources, streamlining access to the Teradata Vantage platform and enabling agents to query, reason, and act with added precision, contextual awareness, and security.

Sumeet Arora, Chief Product Officer at Teradata, commented on the launch:

"AgentBuilder represents meaningful progress in advancing agentic AI for the autonomous enterprise. By combining the flexibility of open-source frameworks with Teradata's AI and knowledge platform and our MCP Server, which provides deep semantic access to enterprise data, we're helping organizations build intelligent agents that are not only autonomous and scalable, but also deeply aligned with their business goals, governance standards, and domain expertise."

Arora further explained the potential impact of this combination:

"Add to that our seamless support across cloud and on-premises environments, and we're delivering a level of flexibility, integration, and contextual intelligence that sets Teradata apart. This is not just about data - it's about delivering trusted, transparent, and complete knowledge to power the next generation of AI."

Support for open-source frameworks

In its initial release, AgentBuilder supports popular open-source frameworks such as Flowise and CrewAI, with integration for LangChain and LangGraph planned for future updates. These frameworks offer modular tools for creating agent workflows, memory, reasoning, and coordination, providing flexibility required for deploying autonomous systems. The integration of these frameworks with Teradata's hybrid infrastructure and governance features is aimed at simplifying real-world adoption of agentic AI applications.

Example Teradata Agents

AgentBuilder includes several pre-built agents designed for key business tasks:

The Teradata SQL agent is designed to interpret natural language questions or requests and convert them into SQL queries for execution on Teradata data warehouse tables. It can identify database schemas, understand table definitions, and optimise queries, and is compatible with open-source multi-agent systems.

The Teradata data science agent functions as a technical agent capable of generating complete machine learning pipelines from natural language requests. By combining large language models with MCP tools and linguistic analysis, it handles multi-step reasoning, contextual understanding, and workflow execution to provide actionable reports and insights.

The Teradata monitoring agent serves as an automation tool for continuous monitoring and management of Teradata databases, servers, and subsystems. Drawing on real-time telemetry data, it maintains system health by detecting anomalies proactively and optimising performance throughout the enterprise.

Focus on governance and scalability

AgentBuilder's inclusion of open-source support and governance features is intended to ensure agents are secure and aligned with enterprise standards, addressing key security and compliance requirements. These capabilities aim to help organisations move beyond pilot projects to fully managed production use of agentic AI across both cloud and on-premises infrastructure.

AgentBuilder will be launched in a private preview in the fourth quarter of 2025, with further details expected as the offering moves towards broader availability.