How-toShadow AIGetting Started
How to build an AI tool inventory in a week
A step-by-step way to find out which AI tools your business actually uses, without a big project or making anyone feel watched.
Phil Evans
Co-Founder & CEO
- Published
- Last reviewed
An AI inventory is a simple list of the AI tools and features in use across your business. It’s the foundation of everything else in AI governance: you can’t set rules, assess risk or train people well until you know what’s there.
This method takes about four hours of focused effort, spread across a week. You don’t need specialist tools.
What you’ll end up with
A single list showing each AI tool, who uses it, what it’s used for, what data goes into it, and a first view of risk. It won’t be perfect. It will be far better than not knowing.
- Time
- 4 hr
- Level
- Beginner
- Steps
- 7
Before you start
- A named owner for the inventory
- Last 6 to 12 months of software and expense records
- Agreement from leadership to frame this as fact-finding, not enforcement
Tools you need
- A spreadsheet or shared list
- A short staff survey form
- Access to your software admin consoles
1Set up your inventory template
Create a spreadsheet with these columns: tool name, vendor, how it’s accessed (free, paid, built into existing software), teams using it, typical uses, data entered, approved (yes, no, unknown), and notes.
Keep it simple. You can add detail later.
2Check what you already pay for
Review software subscriptions and expense claims from the last 6 to 12 months. Look for AI tools by name, and for subscriptions that have recently added AI features.
Add everything you find to the list, even if you’re not sure it’s still in use.
3Look inside your existing software
Many AI features arrive inside tools you already use, such as office suites, CRM, video calls and document management. Log into admin consoles and check which AI features are available and switched on.
Note whether they’re on by default and whether you can control them.
4Ask your people
Send a short, friendly survey. Ask which AI tools people use for work, what they use them for, and whether they ever enter client or personal information.
Say clearly that this is about understanding, not catching anyone out. Allow anonymous answers if it helps. You’ll get more honest responses.
5Talk to team leads
Spend 15 minutes with each team lead. Ask what AI their team relies on, what they’d like to use, and what worries them. These conversations often surface tools the survey misses.
6Add a first view of risk
For each tool, give a simple rating. Higher: client or personal data goes in, output goes to clients, or AI influences decisions about people. Medium: internal information, output reviewed before use. Lower: no sensitive data, low-stakes tasks.
This helps you decide where to focus first.
7Agree next steps and a review date
Share the inventory with your accountable lead. Agree which tools to approve, which to review, and which to stop. Set a date to refresh the list, monthly or quarterly works for most firms.
New tools appear all the time. The inventory is a living document.
Frequently asked questions
What counts as an AI tool?
Any tool or feature that uses AI to generate, summarise, predict or decide. That includes standalone chatbots, AI features in existing software, browser extensions and AI built into specialist systems.
How often should we update the inventory?
Monthly or quarterly. Also update it whenever someone requests a new tool or a vendor adds AI features.
