AI spend intelligence

See what your company is doing with AI.

Really

Depthdata connects ChatGPT, Claude, Copilot, and every other AI tool your teams use into one governed view of adoption, spend, and cost per outcome: evidence your CEO, CFO, CIO, and board can all agree on.

GET EARLY DEMO → SEE THE PRODUCT ↓
READ-ONLY ADMIN APIS · NO AGENTS ON LAPTOPS · PROMPTS NEVER READ
Depthdata
Depthdata
Complete visibility
Every AI tool, seat, and workspace in one governed view instead of five admin consoles
Spend you can justify
Cost per active user, waste from idle seats, and model mix. Finance-grade, not vendor-reported.
Adoption that compounds
See who's getting real value, who's stuck, and route coaching where it moves the number.
Answers for the board
One health score, defensible trends, and an export that drops straight into the deck.
Integrations

Connects to the tools you already run

Read-only connections to ChatGPT Enterprise, Claude, Copilot, Gemini, Cursor, and more. No agents to install, no workflow changes, nothing your employees have to do differently.
THE PRODUCT

Four surfaces.
Zero guesswork.

Depthdata normalizes usage, spend, and outcomes from every admin console into one governed data model, then turns it into four surfaces your leadership team will actually use every week.
01 / Overview

Workspace health, one score

Adoption, depth, and spend across every tool roll up into a single workspace health score, with the KPIs behind it one click away. The first screen your exec team checks on Monday, and the last word in “how’s AI going?”
HEALTH SCORE
KPI DELTAS
12-WK TRENDS
ACTIVE USERS
218/ 287
12%
vs last 30d
TOKENS
24.8M/mo
8%
vs last 30d
AI SPEND
$48.2k/mo
14%
watch · trending up
AVG AI SCORE
72/100
3 pts
DERIVED · published weights
ADOPTION SCORE
Workspace health
0 100
76
/ 100
STRONG
218Active seats
69Dormant seats
76% of licensed seats active weekly
6%
02 / Coaching

Nudges that close the gap

Depthdata identifies what your strongest AI users do differently, packages it into playbooks, and routes nudges to the teams leaving the most value on the table. Detected → nudged → tracked, so enablement effort shows up in adoption.
PATTERN DETECTION
PLAYBOOKS
USAGE PATTERNS
Coaching insights: what to act on this week
03 / Analytics

Trends you can defend

Slice adoption, depth, and spend by department, tool, or seniority. Every number carries a confidence label and an open methodology, so when it lands in the board deck it survives the follow-up questions.
DEPT SLICING
CONFIDENCE LABELS
EXPORTS
TREND · LAST 12 WEEKS
Adoption & tokens
Active users
Tokens
Both lines are still climbing, and token usage is rising faster than headcount. Each active user is doing more over time, so depth is compounding, not just spreading.
+47%ACTIVE USERS · 12WK
1.8×TOKENS/USER VS W1
+58/wkRUN-RATE, ACCELERATING
04 / Leaderboard

Competition that
rewards depth

Department and individual leaderboards score depth of use (chained workflows, tool breadth, retry patterns), not raw message counts. The only way to game it is to actually get better at AI.
DEPTH SCORING
TEAM VS TEAM
WEEKLY MOVERS
Adoption vs. spend by department

Every data
counts.

How it works

From five dashboards to one source of truth

Depthdata is read-only by design. Connect your AI admin consoles and get a governed, historical record of your entire AI footprint, without touching an employee device or reading a single prompt.

Connect your consoles

OAuth into ChatGPT Enterprise, Claude, Copilot, and the rest. Read-only API scopes: nothing to install, nothing your team notices.

One governed model

Usage, seats, and spend normalized into a single schema. “Active user” finally means the same thing across every tool.

Set your baselines

Define what good looks like per department. Depthdata tracks lift against your baseline, not a vendor’s benchmark.

Route the insights

Health scores to execs, nudges to team leads, spend alerts to finance. Each stakeholder gets their view, automatically.

Export anywhere

Push to your BI stack, or export board-ready reports with methodology attached. Your data stays yours.

CONNECTORS / 04

01 ChatGPT Enterprise LLM SUITE
02 Claude LLM SUITE
03 Microsoft Copilot PRODUCTIVITY
04 Gemini LLM SUITE
05 Cursor ENGINEERING
06 Perplexity RESEARCH
07 Notion AI · Slack AI WORKSPACE
08 Mistral · DeepSeek · Grok · GLM LLM

16 tools. One ledger.

Read-only OAuth into the admin consoles you already have. No agents on laptops, no browser extensions, nothing for IT to deploy.

NEW TOOL AT THE COMPANY?

Connected in minutes. Depthdata normalizes its usage into the same adoption, depth, and value model as everything else and history intact.

CONNECTOR TRANSPARENCY

We can only see what vendors expose.
Here is exactly that.

Every Depthdata connector uses the vendor’s official admin APIs with read-only scopes. No agents, no browser extensions, no scraping. This matrix is kept current as vendors ship new APIs, and where a vendor exposes nothing, we say so.

CONNECTOR
SEATS & ROLES
USAGE & ADOPTION
SPEND
AUDIT EVENTS
ChatGPT Enterprise
API
EXPORTS + LOGS
API
API
Claude Enterprise
API
API
API
API
GitHub Copilot
API
API
API
API¹
Gemini (Workspace)
API
API
CLOUD BILLING
API
Vercel + AI Gateway
API
API
API
STREAMING
Notion AI
API
EXPORTS
NOT EXPOSED
STREAMING
Replit
SCIM
EXPORTS
NOT EXPOSED
STREAMING
Lovable
SCIM
NOT EXPOSED
NOT EXPOSED
NOT EXPOSED
OFFICIAL ADMIN API VENDOR EXPORT OR LOG STREAMING WE INGEST NOT EXPOSED YET. THE MOMENT THEY SHIP IT, WE PULL IT

MOST VENDOR ADMIN APIS REQUIRE THAT VENDOR’S ENTERPRISE OR BUSINESS TIER. COVERAGE REFLECTS YOUR PLAN WITH EACH VENDOR. · ¹ REQUIRES GITHUB ENTERPRISE CLOUD.

PLATFORM / 05

Built for the whole org chart.

From individual profiles to board-ready reports, with privacy guarantees that make the whole thing workable.

Platform cards: employee profiles, reports, integrations, security, privacy
EMPLOYEE PROFILES
JM
Jordan Miles
PRODUCT DESIGNER · DEPTH 92
ADOPTION
94
DEPTH
88
BREADTH
76
SAFETY
98

Every person, every tool, one page

Score breakdowns, 12-week trends, tool mix, and coaching history, visible to the employee first.

REPORTS

Board-ready in one click

Branded PDF reports with the same defensible numbers: adoption, spend, department breakdowns, methodology included.

INTEGRATIONS
GPT
CL
GEM
CO
CUR
PX
NO
SL
MI
DS
GR
GLM
LL
API
+

16 tools, zero installs

Read-only OAuth into admin consoles. New tool? Connected in minutes.

SECURITY
SOC 2 TYPE II GDPR SSO / SAML SCIM DATA RESIDENCY AUDIT LOG

Enterprise-grade by default

Role-based access, full audit trail, and your data never trains anyone’s models.

PRIVACY
Share profile with manager
Appear on leaderboard
Share prompt patterns

Consent-first, always

Metadata only, never prompt contents. Employees see and control exactly what’s shared.

COST / ACTIVE USER
$221
 /mo

AI spend ÷ active seats, a measured ratio

FIRST SCAN
31
%

of licenses found idle in the average first scan

WITH COACHING
6
 wks

to close the adoption gap between teams

SETUP
1
 hr

to go live, admin consoles only, zero installs

✦ OPEN METHODOLOGY

Every number, traceable to its source.

No black box. This is the exact pipeline between your vendors’ admin consoles and your board deck. It ships with the product and with every export, so your finance team can audit the math, not take our word for it.

GPT CL GEM CO +12 DAILY SYNC DEPTHDATA SEATS SESSIONS USAGE SPEND READ-ONLY PROMPTS OPEN
TOKENS CREDITS ACTIONS PREM. REQ UNIFIED SCHEMA 1 OF 1 UNIT SESSION ACTIVE USER = 1 DEFINITION ALL TOOLS SPEND OPEN
YOUR FIRST 4 WEEKS W1 W4 W12 W24 BASELINE LIFT PER DEPARTMENT ENG SALES OPS INDUSTRY BENCHMARK VENDOR CASE STUDY OPEN
USAGE SIGNALS TASK SAMPLES P90 P10 6.4 HRS/WK ± 1.2 POINT ESTIMATE SHIPS WITH BAND OPEN
ADOPTION ×0.40 DEPTH ×0.35 EFFICIENCY ×0.25 72 WORKSPACE HEALTH WEIGHTS PUBLISHED · DOCS + EXPORTS OPEN

✦ FAQ

The questions we get asked first.

✦ CONTACT

Know what your AI is actually worth.

Connect your tools once. Get answers you can take to the board.

READ-ONLY ADMIN APIS · NO AGENTS ON LAPTOPS · PROMPTS NEVER READ

GET EARLY DEMO → SEE THE PRODUCT

READ-ONLY ADMIN APIS · NO AGENTS ON LAPTOPS · PROMPTS NEVER READ