Cost agent

The Cost Agent helps data teams identify and reduce warehouse costs — both storage waste and compute runtime. It combines a conversational AI agent with a structured Data tab that surfaces your top optimization candidates, so you can explore interactively, browse a prioritized report, or query the agent through the MCP toolkit.

Capabilities

The Cost Agent covers three areas of cost analysis:

Storage cost analysis

Identifies tables that are wasting storage spend by classifying them into waste patterns, including:

• Unread tables — tables that exist but are never queried by downstream consumers
• Write-only tables — tables that are written to but never read
• Dead-end tables — tables with no downstream lineage (no consumers)
• Static waste — tables whose data has not changed in a significant period
• Zombie tables — tables that are no longer being written to but still consume storage
• Other stale — tables that show other signs of disuse

The agent uses Monte Carlo lineage data to verify that tables are truly unused before recommending cleanup, and estimates the storage savings from removing identified waste.

Compute analysis

Analyzes query groups across your warehouse to surface runtime and performance optimization opportunities. The agent can:

• Rank query groups by runtime, duration, or frequency
• Identify query performance patterns and runtime anomalies
• Surface opportunities to optimize expensive query patterns
• Profile query group performance across time periods

Warehouse usage analysis

For Snowflake environments, the agent provides warehouse-level analysis:

• Ranks warehouses by credit consumption
• Identifies over-provisioned or under-utilized warehouses
• Advises on right-sizing warehouse configurations

Data tab

The Data tab is a report view that surfaces your top storage optimization candidates ranked by size, waste pattern, and risk. Each row shows a risk tier (Low / Needs review / High risk) and downstream dependency count so you know what's safe to act on. Filter by database, schema, risk level, or category. Click Explain with AI on any row to jump into the agent with full context preloaded, or Export the full list to act on it outside the product.

Warehouse support

Accessing the Cost Agent

In the Monte Carlo UI

The Cost Agent is available within the Monte Carlo platform at Agentic Operations > Cost, or through the Operations Agent by asking a cost-related question in natural language. You can ask:

• "Which query groups consume the most runtime?"
• "Which tables are wasting storage?"
• "Show me unused tables in the analytics database"
• "How much could I save by cleaning up zombie tables?"

The agent stays focused on cost-related questions and will redirect out-of-scope requests to the appropriate resource.

Via the Agent Toolkit (MCP)

The storage-cost-analysis skill in the Agent Toolkit exposes storage cost analysis through AI coding agents like Claude Code and Cursor. The performance-diagnosis skill covers query and runtime analysis the same way.

• "What's costing us the most in storage?"
• "Identify unread tables in my warehouse"
• "Estimate cleanup savings for unused tables"

Scheduled cost reports

Set the agent to deliver recurring digests of your storage and query-group analysis on a schedule, so teams can track waste and runtime trends over time without re-running an investigation each time.

Security & data privacy

The Cost Agent uses only metadata, lineage, and query log data — it does not access or process the actual data in your tables. For more detail on security and data privacy, see the AI Features and Technical Information documentation.