AI GovernancePDF + CSVFree

Shadow AI Discovery Worksheet

Shadow AI differs from ordinary shadow IT in what leaves the building: unstructured text that may contain anything, pasted by someone who did not consider it a data transfer. This worksheet is a repeatable discovery sweep.

What you get

  • Four discovery methods with the specific signals to look for
  • A findings register scoring exposure rather than tool popularity
  • A triage path: adopt, replace, restrict, or block
  • A re-run cadence, because this is a recurring sweep not a project

How to use this template

  1. 1

    Start with OAuth grants. Check which third-party applications have been granted access to your M365 or Google tenant. AI tools frequently request broad mail and file scopes, and users grant them without realising what they have authorised.

  2. 2

    Score by what data was exposed, not by user count. A single person pasting customer contracts into a consumer chatbot matters more than forty people generating social images. Rank by data sensitivity.

  3. 3

    Lead with adoption, not blocking. If people are using a tool, it is solving something. Blocking without an alternative moves them to a personal device. Prefer adopting a sanctioned equivalent and blocking the unsanctioned one together.

What's inside

Here's a preview. Unlock the free download to get all 3 sections (2 more below).

Discovery sweep

Run all four. Each catches a different population; relying on one leaves gaps.

MethodWhere to lookWhat to look forRun onFindings
OAuth grantsIdentity provider enterprise applicationsThird-party apps with mail, file, or calendar scopes
Expense and card dataLast 12 months of corporate card and expense claimsRecurring charges to AI vendors; small monthly amounts

2 more sections in the full template

Plus the branded PDF and editable CSV.

Unlock the free download

Related templates

More free templates IT teams use alongside this one.

AI Governance

AI Tool Inventory

You cannot govern AI usage you cannot see, and most AI adoption happens through free tiers and browser extensions that never touch procurement. This inventory is built to surface those, which a record of what you bought will not show.

Get template
AI Governance

AI Acceptable Use Policy

Banning AI tools does not work. It moves usage onto personal accounts where you have no visibility at all. This policy assumes people will use AI and makes the boundary explicit: which tools, which data, and how to get something new approved.

Get template
AI Governance

AI Tool Approval & Intake Form

An AI approval process that takes six weeks guarantees people stop asking. This form is deliberately short: enough to make a sound decision, fast enough that requesting approval stays easier than going around it.

Get template

Frequently asked questions

How do you discover shadow AI usage?

+

Four complementary methods: OAuth grants in your identity provider, expense and corporate card data, browser extension inventories from endpoint management, and network egress logs to AI domains. Each catches a different population.

Why is shadow AI riskier than other shadow IT?

+

Because the data leaving is unstructured free text pasted by someone who did not experience it as a data transfer. A single paste can contain contract terms, customer details, or credentials, and consumer tiers frequently retain and train on it.

Should unsanctioned AI tools be blocked immediately?

+

Only alongside a sanctioned alternative. Blocking without a replacement moves the activity to personal devices where you have no visibility at all. Announce the approved tool first, then block, and treat the discovery as information rather than misconduct.

Shadow AI Discovery Worksheet (Free Template) | ITCommons