AI Agent & Gen AI
Generative AI in Ataccama ONE automates data quality and governance workflows. At its core is the ONE AI Agent, a goal-based assistant that plans and executes complex, multi-step tasks by calling tools in succession.
Beyond the agent, embedded Gen AI features handle everyday tasks like generating SQL or writing descriptions, while the MCP Trust Layer allows you to connect external AI tools like Claude and ChatGPT to your governed data.
All Gen AI features provide suggestions that are easy to review, edit, and refine. Your data and metadata are never used for model training.
What Gen AI enables
-
Intelligent task automation: The AI Agent creates and executes plans for complex tasks like bulk rule creation, catalog management, and data exploration.
-
Less manual work: Automate repetitive tasks like generating DQ rules, writing descriptions, and debugging expressions.
-
Natural language interaction: Use plain-language prompts to generate SQL queries, create DQ rules, and explore your data.
-
Extended data trust: Connect external AI tools to your governed data through the MCP Trust Layer, so AI assistants can answer questions about your assets, data quality, and ownership.
Get started with Gen AI
Learn prompting best practices
Write clear, specific prompts to get better results. Include context, provide examples of expected formats, and iterate on outputs.
Work with AI Agent
Give the AI Agent a goal and let it develop and execute an action plan. Use it for creating DQ rules, data quality assessments, data exploration, or documentation questions.
See AI Agent.
Use embedded Gen AI features
Generate SQL from plain language, create DQ rules from descriptions, auto-generate catalog descriptions, and more. See Embedded Gen AI Features.
Connect external AI tools
Use the MCP Trust Layer to give tools like Claude, ChatGPT, and Copilot read-only access to catalog metadata, business terms, and quality metrics — respecting your platform permissions. See MCP Trust Layer for External AI Agents.
Learn more
-
AI Governance and Security: How Ataccama approaches data privacy, data handling, and responsible AI practices.
-
Install MCP Trust Layer: Set up connections for external AI tools.
Was this page useful?