AI Usage

Know what AI costs. See how it’s used.

Your AI tools are becoming part of everyone’s day. June brings spend, adoption, and efficiency into one clear view.

app.juneops.com/ai

June

Agent

Requests

Reachouts

Signals8

Access

Policies

Intelligence

Runbooks

Knowledge Base

Resources

Assets

Software

AI usage

Claude

Cursor

OpenAI

Employees

Groups

Departments

Locations

AI usage

Total spend$14,400

Tokens12B

Active tools3

People with usage120

Departments6

Shared API spend$2,000

Cumulative spend

A single view across your AI tools

By provider

ClaudeCursorOpenAI

Spend by provider

Tokens / spend

Claude6B$8,400

Cursor4B$3,600

OpenAI2B$2,400

$12,400 employee · $2,000 shared API

Spend by department

Employee spend

Engineering$6,200

Product$2,480

Design$1,240

Operations$992

Sales$868

Finance$620

Claude

Cursor

OpenAI

Claude, Cursor, and OpenAI in one view.

Bring Claude, Cursor, and OpenAI into one view. Follow spend across providers, departments, and the people using them.

Real usage, not just seats.

Go beyond assigned seats. Understand active use, popular workflows, and where a team could use a little help getting started.

What’s driving the bill?

Explore model mix, caching, and cost trends. See what’s driving the bill before deciding where to make a change.

The details behind the total

Every dollar traced to a person and a model

Follow any number back to its source. Providers, departments, people, models, all the way down. So when finance asks, you already have the answer.

Claude Connected

Anthropic · Usage and billing

Aug 1–31, 2026

Employee spend$7,200Attributed to 80 active people

Shared API spend$1,2003 organization sources

Billed spend$8,4005.6% vs. July

Annual commitment$120,000$68,000 used · Jan–Dec 2026

Cumulative spend

Across all Claude models

6B tokens

Claude SonnetClaude OpusClaude Haiku

Spend by model

Claude Sonnet3B tokens$4,200

Claude Opus1B tokens$3,000

Claude Haiku2B tokens$1,200

Spend by department

Employee spend

Engineering$3,600

Product$1,440

Design$720

Operations$576

Sales$504

Finance$360

You’re on track. There’s still room to spend smarter.

At August’s pace, you’ll finish the year at $101,600 against a $120,000 commitment. Two checks could reduce usage costs by about $650/month.

Within budget

A clearer view of the year ahead

Based on August’s $8,400 monthly pace

Jan–Dec 2026

Actual spendProjected at August’s paceWith opportunities reviewed

Worth a closer look 2

Potential monthly reduction

16 seats haven’t been used this month$320

Check in with the employees before changing access. Some may be on leave or waiting to get started. Reassigning unneeded seats could reduce future costs.

Suggested owner · IT Operations

Long contexts account for 18% of spend$330

Review repeated context in high-volume workflows with the engineering team. Shorter prompts and reusable context may lower costs without changing the task or model.

Suggested owner · Morgan Wu

Looking good · Cache usage, model mix, skill costs, and spend caps

Opportunities are estimates, not guaranteed savings. Reviewing both would bring the remaining four-month forecast to $99,000 for the year.

People with activity83.3%80 of 96 available seats

Daily active users52Average · Aug 25–31

Sessions per person241,920 Code + Cowork sessions

Inactive seats16Ready for a usage review

Who’s making AI part of their day?

Daily active people · Final week of August

Product adoption

People can use more than one product

Product	Active people	Daily users¹	Monthly activity	Per person
Chat	64	46	2,560 messages	40 messages
Claude Code	48	34	1,680 sessions	35 sessions
Cowork	16	10	240 sessions	15 sessions

¹ Average daily users in the final week. Product audiences overlap.

Employee cost / person$90$7,200 / 80 active people

Effective cost / M tokens$1.40Across all billed usage

Cache hit rate91.6%Of eligible input tokens

Tokens / request85.7k6B tokens / 70k requests

Understand the mix behind your spend

Effective costs reflect your observed usage, including caching

Model	Spend	Tokens	Requests	Cost / M tokens
Claude Sonnet	$4,200	3B	40k	$1.40
Claude Opus	$3,000	1B	10k	$3.00
Claude Haiku	$1,200	2B	20k	$0.60
All models	$8,400	6B	70k	$1.40

Caching is doing its job.

91.6% of eligible input tokens are served from cache. Keep that context when comparing model costs.

Use cost as a starting point.

Different work needs different models. Review the task and outcome with the team before making a change.

See which shared workflows are catching on.

Compare adoption and attributed cost across published and personal skills.

Published skills

3 skills · 840 uses · $420 attributed cost

Skill	People	Uses	Cost	Per use
pull-request-reviewA consistent first pass on every change	32	360	$180	$0.50
release-readinessCheck the details before a release	24	240	$144	$0.60
customer-briefTurn account context into a useful brief	20	240	$96	$0.40

Personal skills

2 skills · 90 uses · $42 attributed cost

Skill	People	Uses	Cost	Per use
research-summaryAlex Kim	1	60	$24	$0.40
design-handoffJordan Lee	1	30	$18	$0.60

Skill costs are part of your existing Claude spend, not an additional charge.

Active people80Across 6 departments

Employee spend$7,200Shared API usage excluded

Employee tokens5.2B

Average per person$90

Employee usage

Employee	Tokens	Sessions	Spend	Monthly cap
Morgan WuEngineering	400M	144	$480	$480 / $600
EPEthan ParkEngineering	350M	126	$420	$420 / $600
Maya ChenEngineering	300M	108	$360	$360 / $600
Alex KimProduct	200M	72	$240	$240 / $600
Jordan LeeDesign	160M	58	$192	$192 / $600
NBNora BennettDesign	150M	54	$180	$180 / $600

Showing 6 employees · 80 active this month

Use individual usage to support a conversation about tools and budgets. Activity alone doesn’t measure productivity.

Keep shared usage in the picture.

API projects and shared services stay visible, even when their costs don’t belong to one employee.

Organization sources3

Shared API spend$1,200

Shared API tokens800M

Organization usage

Shared sources included in billed spend

Source	Attribution	Tokens	Spend
Production APICustomer-facing services	Organization	400M	$620
Internal toolsShared team workflows	Organization	250M	$380
SandboxDevelopment and evaluation	Organization	150M	$200
Total	3 sources	800M	$1,200

Every dollar has a place. $7,200 employee + $1,200 shared API = $8,400 billed.

Usage and billing connectedAugust 2026

Put the follow-through on a schedule

Review AI spend every month with a June runbook.

Set the review once. June forecasts AI spend, compares it with your contract value, and keeps your team ahead of budget surprises.

Meet Runbooks

Runbook

Claude spend forecast

Active

Monthly · First weekday, 9:00 AM

What June checks

See where Claude spend is heading and whether it fits within the contract.

Review recent spend and changes in the daily spend rate.
Forecast spend for the next month.
When a contract is available, compare the forecast through its end date with the contract value.
Send the IT team a short summary of the outlook.

Latest forecast · August 2026Completed

Claude spend is on track.

Contract value Jan–Dec 2026: $120,000
Spent so far Through August: $68,000
Forecast at contract end: $101,600

$18,400 below contract valueProjected at August’s pace of $8,400/month.

Frequently asked

Which AI tools can I connect?

June connects with Claude, Cursor, and OpenAI. The available views depend on the provider, your plan, and the usage data its connector supplies. The interactive example above shows the Claude experience.

How does June separate employee and shared usage?

Usage that can be attributed to a person appears in employee views. Shared API projects and other organization sources are shown separately, so you can understand both without assigning every dollar to an employee.

Are all the cost figures billed amounts?

June distinguishes billed spend from attributed or estimated costs where the provider makes those distinctions. Effective model costs reflect actual usage patterns, including caching; they aren’t a vendor’s list price.

Can I see adoption and skill usage?

Where supported by the provider, June shows active people, product activity, and skill usage. For Claude, you can compare published and personal skills, including how often they’re used and their attributed cost.

Can June review AI spend on a schedule?

Yes. Use a Runbook to describe what to review, choose a schedule, and specify the findings or follow-ups you want. Start with a monthly budget review or a check for seats that haven’t been used.

Get a handle on what AI costs.

See AI Usage in action