Find out what your AI costs, and where it is exposed.
Upload your usage data. A first read in your browser now; see potential savings and the signals worth a closer look.
Three steps. No software to install.
Export usage data from the tools you already pay for. We show you how, for each one.
Upload it here. Your browser gives you a first read on the spot. Nothing leaves your machine at this step.
Review cost upside, personal-account signals and tools outside your approved list. Then decide whether to follow up with us.
Pick a source. Follow three steps.
- Open Workspace settings → Workspace analytics.
- Select Export data, then the Users data type and a date range.
- Generate CSV. Keep one reporting cadence and non-overlapping periods.
Columns we look for: cadence · period_start · period_end · account_id · public_id · messages · email · model_to_messages. Model counts allow a routing scenario; API-equivalent costs are estimates, not your ChatGPT subscription bill.
OpenAI export reference- Open the Microsoft 365 admin centre.
- Go to Reports → Usage → Microsoft 365 Copilot; select a reporting period.
- Export the user-detail CSV. Use one report window; v2 adds prompt and active-day counts.
Columns we look for: Report Refresh Date · Report Period · User Principal Name · Copilot app Last Activity Date. Shows account-domain signals and inactive-user candidates; this export does not contain model names or subscription costs.
Microsoft CSV reference- Open your Anthropic organisation's API usage reporting.
- Export token usage by model and workspace using the documented Admin API cookbook.
- Choose its usage CSV here. For a different Console layout, use the generic template below.
Columns we look for: date · model · service_tier · workspace_id · uncached_input_tokens · output_tokens · cache_creation_tokens · cache_read_tokens. Supports a model-cost scenario. Workspaces are not individual users.
Anthropic CSV reference- Open Admin console → Generative AI → Gemini reports.
- Select User-level usage for the period shown.
- Download the table or usage counts. Confirm the source when this page asks.
Illustrative English layout: Email · Overall usage · Active days · Gemini app · Gmail · Docs. Google's guide describes fields rather than stable CSV headers. Usage levels support adoption checks; they do not reveal model names or token costs.
Google export guidance- Export a statement or expense report for the period you want to review.
- Filter it to relevant AI merchants and save as CSV.
- Use date, merchant and amount columns; add currency and claimant where available.
Columns we look for: date · merchant · amount · currency · claimant/cardholder · cost centre. Amounts and refunds support a consolidation scenario. Transactions alone do not prove actual product use.
- Export application activity or sign-in records from your monitoring tool.
- Filter to relevant applications and choose a period.
- Use date, user and application columns; add sign_in_count where available.
Columns we look for: date/Date (UTC) · user/User principal name · application · sign_in_count. Highlights domain signals and tools outside your list. Sign-in attempts do not establish spend or data exposure.
Other sources: use the generic usage CSV template. Add model and input/output token counts or recorded costs for a stronger savings estimate. Extra columns are ignored.
Drop your file here for a first read.
CSV or TSV exports, up to 10 MB. Save Excel sheets as CSV first. We read usage fields, not prompts or documents.
See what your first read could reveal.
No file needed. Choose the example closest to your organisation and explore the report.
Usage only. Never content.
Who used which tool, when, how much, and what it cost where supplied. Email domains and model usage help highlight the opportunities worth reviewing.
No prompts, no outputs, no documents, no customer content. Remove these before choosing a file; unrelated columns do not contribute to your report.
Names and email addresses are not shown individually in the report. Only domain counts and usage summaries reach this page from its local worker.
Your file is processed locally. No file, email address or report is sent to GenWA. Follow-up sharing will use a separate facility when available.
This check makes no AI API calls. Your file is not pooled, sold or sent to a model. Optional website analytics are managed separately in Cookie settings.
Your summary stays in page memory, not browser storage. Clear analysis or leave the page to discard it. Agree the data-handling terms before any future sharing.
View the previous page for comparison.
Find out where your AI stands. Then decide.
Under a minute, an upload, or a visit. Something for the board either way.