Claude Code + Usage Intelligence Agent: See How Your Data Tools Are Actually Used

The Claude Code usage intelligence agent is an MCP server from Data Workers that shows which data tools, MCP servers, dashboards, and tables are actually used across your organization. From your terminal, it surfaces adoption metrics, abandoned assets, and tool effectiveness — so you stop maintaining things nobody touches.

The Claude Code usage intelligence agent shows you exactly how your data tools, MCP servers, and data assets are being used across your organization — all from your terminal. If you have ever wondered whether anyone actually uses the dashboard you built last month, or which MCP tools your team relies on most, or which data tables are critical versus abandoned, the usage intelligence agent from Data Workers provides the answers. It is an MCP server that gives Claude Code visibility into adoption metrics, usage patterns, and tool effectiveness across your entire data stack.

Data teams make infrastructure decisions — which tools to invest in, which tables to maintain, which pipelines to deprecate — based on assumptions about usage. Those assumptions are usually wrong. The table you think nobody uses powers a critical executive report. The tool you invested heavily in has three active users. The usage intelligence agent replaces assumptions with data, so you make infrastructure decisions based on facts.

The Visibility Gap in Data Infrastructure

Most data teams have no clear picture of how their tools and data are actually used. Warehouses track query history, but parsing that into meaningful usage metrics requires custom analysis. BI tools have their own analytics, but they do not tell you about direct SQL access. MCP servers handle tool calls, but aggregating usage across agents requires additional instrumentation.

This visibility gap leads to bad decisions: maintaining expensive tables that nobody queries, building features for tools that nobody uses, deprecating assets that turn out to be critical to one important but quiet consumer. The usage intelligence agent closes this gap by aggregating usage signals from across your stack into a unified view.

Querying Usage Metrics from Your Terminal

With the usage intelligence agent connected to Claude Code, you can ask usage questions directly:

claude "Which MCP tools are most used by our team and which are never used?"

The agent returns a ranked view of tool usage:

This view immediately tells you which tools are delivering value, which are underutilized (and might need better documentation or training), and which can be safely deprecated.

Data Asset Usage Analysis

Beyond tool usage, the agent analyzes how your data assets are consumed:

claude "Which tables in the analytics schema are queried most and which are candidates for deprecation?"

• •High-value tables: fct_revenue (queried 2,400 times/month by 15 users), dim_customers (1,800 queries, 22 users), fct_orders (1,200 queries, 18 users)
• •Low-usage but critical: dim_fiscal_calendar (45 queries, 3 users — but used in every revenue calculation)
• •Deprecation candidates: tmp_migration_staging (0 queries in 60 days), old_fct_orders_v2 (2 queries in 90 days, both by same user exploring history)
• •Zombie tables: 23 tables with zero queries in the last 90 days, costing $1,200/month in storage

The agent distinguishes between tables that are unused (safe to deprecate) and tables that have low query volume but are critical dependencies (like lookup tables used in every join). This distinction prevents the common mistake of deprecating a table that turns out to be load-bearing.

Team and User-Level Analytics

The usage intelligence agent also provides team-level adoption metrics:

claude "How is each team using Claude Code and the Data Workers agents?"

The agent shows adoption patterns across teams, helping you identify which teams are getting the most value, which teams need onboarding support, and where usage patterns suggest workflow improvements. This data is invaluable for internal champions who need to justify tool investments and demonstrate ROI to leadership.

Before and After: Infrastructure Decision-Making

Custom Usage Dashboards and Reports

For recurring reporting needs, you can ask the agent to generate structured reports:

• •claude "Generate a monthly usage report for Data Workers agents — adoption, top tools, trend lines" — executive-ready summary
• •claude "Which new data assets were created this month and what is their adoption so far?" — new asset tracking
• •claude "Compare tool usage this quarter versus last quarter" — trend analysis for planning
• •claude "What is the estimated time saved by each agent based on usage volume?" — ROI quantification

These reports are generated from live data, not from static dashboards that someone has to maintain. They are always current and always accurate.

Getting Started with Usage Intelligence

The usage intelligence agent aggregates data from your warehouse query history, MCP server logs, BI tool analytics, and Data Workers platform telemetry. Follow the Getting Started guide to install Data Workers and the Claude Code Setup guide to connect the agent. The Docs cover advanced features including custom metric definitions, retention analysis, and integration with product analytics tools.

Start by running a usage audit on your warehouse to discover which tables are truly valuable and which are candidates for cleanup. Visit the Product page to see all 15 agents.

Make infrastructure decisions based on data, not assumptions. Book a demo to see usage intelligence across your data stack.