> For the complete documentation index, see [llms.txt](https://docs.revenium.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend.md).

# AI Assistants

AI Assistants brings connected assistant usage and cost into one place under **Intelligence > AI Assistants**. Review spend by department and person, then open each assistant tab for its own usage and cost detail.

### Connect your assistants

Cursor and GitHub Copilot use organization-wide provider connections under **Integrations > Providers**. Connect your provider first: see [Connecting Providers](/get-started/quickstart/connecting-providers.md). See [Setup Cursor](/track-and-control-costs/analyze-ai-tooling-spend/setup-cursor.md) and [Setup GitHub Copilot](/track-and-control-costs/analyze-ai-tooling-spend/setup-copilot.md). For telemetry setup, see [Setup Claude Code](/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code.md), [Setup Claude Cowork](/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-cowork.md), [Setup Gemini CLI](/track-and-control-costs/analyze-ai-tooling-spend/setup-gemini.md), and [Setup Codex CLI](/track-and-control-costs/analyze-ai-tooling-spend/setup-codex.md).

For individual CLI onboarding and historical backfill, use the [Revenium CLI](https://github.com/revenium/revenium-cli-node). Claude Code and Gemini setup validate the Revenium key, test the connection and try to update your shell profile; Codex setup checks the key format and configures telemetry, and `revenium-codex test` checks the connection. For an organization-wide Claude Code rollout, an administrator configures telemetry once in managed settings, as described in [Setup Claude Code](/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code.md).

Claude Code, Codex and Gemini CLI each have a ready-to-paste setup prompt. In the AI assistant setup, choose the assistant, select **Copy prompt** and paste it into the assistant. Claude Cowork gets step-by-step instructions, and Cursor and GitHub Copilot connect through a provider form. For Claude Code, Claude Cowork, Codex and Gemini CLI the copied text includes your Revenium key, so share it only with the people who set up the assistant. Before you connect anything, the Claude Code, Claude Cowork, Codex, Gemini and Copilot dashboards show sample data, clearly labelled and with a button to connect the assistant.

If a developer's Claude Code usage is missing or looks too low, see [Troubleshooting](/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code.md#troubleshooting).

### Choose a pricing mode

Fixed seats and subscriptions and API-priced traffic have different cost meaning. Configure the distinction per team so Revenium knows whether assistant traffic stays in AI Assistant analytics as API-equivalent estimates or counts as real spend across cost reports. See [Configure AI Assistant Pricing Mode](/track-and-control-costs/analyze-ai-tooling-spend/configure-coding-assistant-billing-models.md).

### Find unused seats

To find the paid seats that go unused on Claude Code, Cursor, Gemini and Codex, and what they cost each month, see [Seats and Licenses](/track-and-control-costs/analyze-ai-tooling-spend/seats-and-licenses.md).

### Overview

The Overview shows person-attributed AI spend, the number of departments and people, and dollars per million tokens for the period you choose. Below the headline figures are Daily AI Spend, Spend by AI Source, Spend Mix and Spend by Department & Source, then department comparisons (top spending, highest cost rate, movers and spend concentration), Shadow Spend by Department and the Departments table. Until [departments are set up](/track-and-control-costs/attribute-spend-to-departments.md), the Overview shows a Set up departments notice with the per-person view below it.

### By Employee

The per-tool dashboards each scope to a single assistant. The **By Employee** tab pivots the same data around people instead, rolling every connected assistant (Claude Code, Claude Cowork, Cursor, Gemini, Codex, Copilot) and direct API spend up to the individual. Where a per-tool view answers "what did Cursor cost," By Employee answers "what did any one person cost across everything": the view you need for chargebacks and for spotting concentrated spend. It needs no extra setup; it draws on the assistants you've already connected.

By Employee shows spend and usage per person, not prompts or code. Revenium does not store code, prompts or conversation content from AI assistants (see [AI Assistant Data Reference](/track-and-control-costs/analyze-ai-tooling-spend/ai-coding-data-reference.md)). For calls you meter through the SDK or API, prompt capture is off unless a team admin turns it on, and viewing captured prompts needs its own permission (see [Debug Logs & Traces](/optimize-performance/debug-logs-and-traces.md)).

Start with the **Attribution Coverage** figure at the top. It tells you how much of your identifiable AI spend has been tied to a known person, and it's the trust indicator for every number below: if coverage is low, the per-employee figures describe only the attributed slice, not your whole AI spend.

From there the page surfaces:

* **Headline KPIs**: Attributed AI Cost, Active Employees, Avg Cost / Employee, Blended $/Mtok (a single efficiency number for the whole population), and the split between assistant and direct-API spend.
* Callouts for spend concentration and cost-rate outliers, when the data supports them.
* **Charts** — daily spend trend, spend by AI source, spend mix, and how cost is distributed across employees. The daily spend trend can also be broken out by vendor.
* **All Employees** table: per-person cost, % of total, tokens and $/Mtok, requests, channel, and last active, exportable to CSV for finance review.

The **Shadow Spend — Account Types** view splits spend into Corporate and Personal. That split is a **heuristic** based on a configurable corporate-domain allowlist, not a verified employee directory. Read it, and the page as a whole, as a cost-governance and coverage tool: it exists to answer "what are we spending, on whom, and how much of it can we see", not to monitor individuals.

Select a bar in a Shadow Spend chart, on **By Employee** or on the AI Assistants **Overview**, to open the people behind it.

#### Filter the view

Subscription-based assistant usage is included by default; use the page filters to exclude it when you want to, and to focus on the provider data and model tiers relevant to the analysis. Filters narrow the current result set before you group, choose columns, or export it.

#### Compare groupings and team benchmarks

Group the table by employee, tool, model, model tier, or model maker. The table adds team-average cost-per-million-token context so an individual result can be compared with the wider team rather than read in isolation.

#### Choose columns and export the result

Select the columns you need for the current analysis. Exporting to CSV uses the current result set, so apply the required filters and grouping before creating the file.

#### Set a budget for an employee

Select **Set a budget** from an employee's detail to open a budget alert with that person prefilled. Review the form and submit it before the rule is created; opening the form does not save it automatically.

For the full rule workflow, see [Guardrails](/track-and-control-costs/set-budgets-and-alerts.md).

***

### Tickets

Tickets shows what AI work costs for each engineering ticket, most of it from coding assistants, so you can use past ticket costs when planning similar work. See [Tickets](/track-and-control-costs/analyze-ai-tooling-spend/tickets.md).

### PR Health

PR Health shows which pull requests need a decision, which were abandoned, and how much coding-assistant spend sits behind them. See [PR Health](/track-and-control-costs/analyze-ai-tooling-spend/pr-health.md).

### Claude Code

The Claude Code tab shows spend, merged pull requests, people, findings, and seat and plan costs. See [Claude Code Dashboard](/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-code/claude-code.md) for how to read those measures.

### Claude Cowork

The Cowork dashboard has **Overview** and **People** tabs. On Overview, review what changed, findings, spend and sessions over time, spend by department or person, and cost by model; use People to inspect individual usage. See [Claude Cowork Dashboard](/track-and-control-costs/analyze-ai-tooling-spend/setup-claude-cowork/claude-cowork.md) for how to read the results.

### Cursor

The Cursor tab shows usage and costs synced from its team API. See [Setup Cursor](/track-and-control-costs/analyze-ai-tooling-spend/setup-cursor.md) for how to connect it and verify the first sync.

### Gemini

The Gemini tab shows Gemini CLI usage and costs from telemetry. See [Setup Gemini CLI](/track-and-control-costs/analyze-ai-tooling-spend/setup-gemini.md) for configuration and connection checks.

### Codex

The Codex tab shows Codex CLI usage and costs from telemetry. See [Setup Codex CLI](/track-and-control-costs/analyze-ai-tooling-spend/setup-codex.md) for configuration, a test metric, and backfill.

### Copilot

The Copilot tab shows GitHub Copilot usage and costs, with Cost by Model for supported dates. See [Setup GitHub Copilot](/track-and-control-costs/analyze-ai-tooling-spend/setup-copilot.md) for the provider connection and refresh timing.

### <i class="fa-wrench">:wrench:</i> Skills

Assistant cost tells you what a tool is spending. **Skills** tells you what it is spending it on, breaking usage down by the skills your assistants actually invoke rather than by the assistant as a whole. It is the view to reach for when a cost figure is understood but the behavior behind it is not.

Filter the view by person to scope it to one developer or a small group, which turns a team-wide distribution into a specific question about specific work. Filters are held in the page address, so the view you are looking at is the view you get back when you reload it, and the link you paste into a thread opens for a colleague exactly as it looked for you.

**Spend in skill sessions** puts the rest of the page in proportion. It is the spend from sessions where a skill ran, shown with the number of those sessions against the total for the workspace. A workspace where skills touch a small share of sessions reads very differently from one where they touch most of them, and this figure tells you which one you have. If a filter narrows the skills counted but not the sessions they are compared against, the page says so above the figure.

The invocation mix separates the two reasons a run has no named trigger. **Inside another skill** means one skill called another, which is ordinary nesting. **Not matched** means the run has no start record, so the telemetry did not arrive. It does not mean nothing happened.

Exports cover every skill matching your filters, in the sort order on screen, rather than only the rows already loaded into the page. The page itself explains how spend is attributed to skills. Skill cost is currently an approximation: use it to compare skills and to see where spend concentrates, not as an exact amount. Skill-labelled spend counts only the calls made in the turn that started a skill, so work that continues in later turns or is handed to a subagent is left out and the labelled figure usually undercounts a skill. **Spend in skill sessions** is usually the upper bound, so the real cost usually sits between the two; a skill call that arrives without a session can make the labelled figure the larger one. Open **How we attribute spend to skills** on the page for the full explanation.

***

### <i class="fa-forward">:forward:</i> What to Do With What You Find

Use the **Departments** table on Overview to see which departments drive AI spend. On By Employee, group by department to see the **API-equivalent** column alongside department cost when subscription-based assistant usage is included.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.revenium.io/track-and-control-costs/analyze-ai-tooling-spend.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.