Teradata Unveils AI Agent System That Slashes Enterprise Data Workflow Costs by 58%

Teradata Introduces Three-Component AI Agent System for Enterprise Data

Teradata has transformed its Tera AI-powered workspace into a comprehensive AI agent system designed to tackle the escalating costs and complexity of agentic execution in enterprise data environments1

1

. The expansion, announced from San Diego, introduces three core components: Tera Context Engine, Tera Harness, and Agent Skills3

3

. Originally launched in May as part of Teradata's Autonomous Knowledge Platform, Tera now addresses a critical pain point that enterprises are discovering as they scale agentic workloads—uncontrolled token consumption from agents that loop endlessly without advancing tasks1

1

.

The system enables business analysts, data engineers, and database administrators to interact with enterprise data through natural language requests instead of writing SQL queries or code2

2

. Users can ask Tera to analyze information, build data pipelines, or manage infrastructure while working within their organization's access rules and access policies2

2

.

Tera Context Engine Connects Data Without Movement

The Tera Context Engine functions as a context and orchestration layer that connects databases, data platforms, and catalogs without requiring companies to move their data between systems3

3

. It brings together metadata, data lineage, business definitions, and governance controls so governed agents can interpret requests within the specific context of the organization using them2

2

. The engine incorporates Industry Knowledge Models containing terminology and relationships specific to different industries, enabling agents to understand sector-specific language and workflows3

3

. Teradata emphasized that the system can trace AI outputs back to their sources and apply consistent policies as information moves between systems, addressing enterprise concerns about data governance in multistep data workflows2

2

.

Tera Harness Reduces Token Usage Through Strategic Execution Planning

Tera Harness serves as the execution layer that determines how agents approach tasks and how workflows are routed1

1

. The system creates an execution plan before sending work to an LLM, batches independent tasks, and drops model or tool calls that do not advance the task1

1

. Built on a Go-native engine to handle multiple concurrent operations, Tera Harness applies 84 execution patterns before inference and limits workflow steps based on progress, reducing repeated LLM reasoning and the token usage associated with unproductive agent loops1

1

. The system includes checkpointing capabilities that allow long-running tasks to pause and resume, enabling recovery from infrastructure failures3

3

. It can also pause for human approval before taking sensitive actions, applying controls before an action runs to limit unauthorized or destructive operations2

2

.

Benchmark Tests Show Significant Cost Reductions in Inference Costs

Teradata provided benchmark results demonstrating substantial improvements in efficiency and inference costs. On the SWE-bench Pro benchmark using the Opus 5 model, Tera used 73% fewer tokens than Claude Code, completed tasks 42% faster, and incurred 58% lower total cost while achieving higher task completion rates1

1

. On the data-eng-bench pipeline engineering benchmark, Teradata reported 53% lower cost per reliably solved task than Snowflake Cortex Code, based on Snowflake's published results2

2

. These benchmark tests reflect Teradata's testing conditions and may not represent performance across all enterprise workloads, the company cautioned2

2

.

Agent Skills Package Reusable Functions for Data Engineering Tasks

Agent Skills are pre-built, reusable functions that package common data engineering, data analysis, and data science tasks3

3

. Platform Agents handle operational work like workload tuning and compute sizing, while Analytics Agents convert natural language requests into SQL and Python code3

3

. Organizations can connect their own tools through the Model Context Protocol, extending the system's capabilities beyond Teradata's built-in functions2

2

. "Most enterprises are not starting from scratch with AI. They are dealing with tools that do not work together and a skills gap that makes those tools hard to use at scale," said Sumeet Arora, Teradata's chief product officer2

2

.

Cost Control Creates Workflow Predictability but Introduces New Tradeoffs

Analysts note that Tera's focus on controlling agentic execution costs addresses a growing enterprise concern. "Tera Harness is attacking the part of agentic AI that enterprises are only now discovering hurts, which is that an agent left to reason its way through every step will happily burn tokens on loops that never move the task forward," said Ashish Chaturvedi, executive research leader at HFS Research1

1

. The ability to constrain agent execution improves workflow predictability, making agentic costs more consistent and budgeting decisions easier, according to Advait Patel, senior site reliability engineer at Broadcom1

1

. However, the approach involves tradeoffs. "If the Harness drops a step it considers unnecessary and that step turns out to matter, the enterprise saves money but gets a worse answer," Patel warned1

1

. Developers may shift from directing individual steps to defining correct outcomes and reviewing agent results, while also maintaining skills, instructions, and guardrails for production operations1

1

.

Deployment Support Through AI Services Launching Q4 2026

Teradata is offering AI Services to help customers identify use cases, configure Industry Knowledge Models, and move enterprise knowledge into production faster2

2

. The services include implementation support for deploying governed context and configuring the business knowledge Tera needs for operation2

2

. Customers will be able to choose their models and run workloads in cloud, on-premises, or sovereign environments2

2

. Tera Context Engine, Tera Harness, and Agent Skills will be available in Q4 2026, though the announcement did not include pricing information3

3

. As enterprises attempt to give AI agents sufficient information and authority to complete work without losing control of data access and approvals, Teradata's system represents a bet that pre-planned execution and built-in governance will become competitive differentiators in the enterprise data market.