17. AI Module Starter Pro Business

Ask questions about your catalog in natural language and get instant, context-aware answers based on your documented assets.

Available on Starter, Pro, and Business plans Usage is measured in tokens (input + output) per month. Each plan includes a monthly token budget shared across all AI features: chatbot, Fill with AI, and Playbook generation. Approximately 50 / 500 / 3,000 average interactions per month on Starter / Pro / Business.

14.1 What You Can Ask

The AI assistant has full context of your catalog: asset names, types, descriptions, owners, tags, sensitivity classifications, and glossary terms. You can ask it questions in plain language, exactly as you would ask a colleague.

Example questions:
"Which tables contain customer data?"
"What is the source of the Monthly Revenue KPI?"
"Show me all assets owned by the Finance team."
"Which tables are marked as confidential?"
"What does the 'churn_rate' column mean?"
"Which reports depend on the sales_orders table?"

14.2 How It Works

When you submit a question, Metaustral builds a context snapshot of your catalog metadata and sends it alongside your question to Claude (Anthropic). The model returns a natural language answer displayed in the chat interface. No raw data from your databases is ever sent — only catalog metadata (names, descriptions, types, owners, tags).

sequenceDiagram participant U as User participant M as Metaustral participant AI as Claude (Anthropic) U->>M: Natural language question M->>M: Build catalog context snapshot M->>AI: Question + catalog metadata AI-->>M: Natural language answer M-->>U: Display answer in chat M->>M: Log usage (quota tracking)

AI module: question → catalog context → Claude → answer displayed in chat

14.3 Monthly Quota

Usage is measured in tokens (input + output) and resets on the 1st of each calendar month. The budget is shared across all users and all AI features in your workspace: chatbot, Fill with AI, and Playbook generation. Monitor consumption from Admin → Plan.

Plan	Monthly token budget	Approx. interactions
Free Free	Not included	—
Starter Starter	250,000	~50
Pro Pro	2,500,000	~500
Business Business	15,000,000	~3,000

Interactions that involve more context — a large catalog, many columns, or a complex lineage graph — consume more tokens. The approximate interaction counts above assume an average of ~5,000 tokens per interaction.