The AI tool inventory
The inventory lists the AI your organization uses: coding tools, AI web apps, models and model deployments, API keys, IDE and browser extensions, MCP servers, agents and built-in AI features.…
Written By Christopher Scaminaci
Last updated 17 days ago
The inventory lists the AI your organization uses: coding tools, AI web apps, models and model deployments, API keys, IDE and browser extensions, MCP servers, agents and built-in AI features. StackJack adds an item the first time it sees it. An administrator then approves or denies it. The inventory is the base of your AI register for clients, auditors and insurers.
At a glance
- What it does: keeps one list of every AI tool, who uses it, and whether it is approved.
- Plan: every plan.
- Who: any governance role reads it. An Administrator approves, denies and edits. The people who use each item are shown to Administrators and Reviewers only.
- Where: AI Governance → AI inventory → Inventory. The page title is "Inventory".
- Needs: a source that reports AI use.
How items get on the list
StackJack adds an item from:
- the activity it receives,
- the admin consoles of the vendors you connected,
- Microsoft 365,
- the browser and network connectors you connected, where they report AI use,
- a request from your staff (see Tool requests),
- you, with Add an AI tool.
What each item records
- Name, kind, vendor, and where it was seen: telemetry, a vendor admin console, Microsoft 365, a browser or network connector, a staff request, or added by hand.
- Who owns it, and the region where its data is processed.
- The contract you hold with the vendor (none, a data processing agreement, a business associate agreement) and whether training on your data is switched off (yes, no, not known).
- The kinds of data it handles, and whether it is used for employment decisions, faces your clients, or processes your clients' data.
- The EU AI Act risk class: not screened, prohibited, high risk, limited risk or minimal risk. Use Suggest a risk class to get a suggestion from AI (Jev), labelled "Suggested by AI (Jev)". It is only a suggestion until a person selects Confirm this class.
- Its approval status.
- When it was first and last seen, how many people used it in the last 30 days, and an estimated cost for the last 30 days when a figure exists. Costs are never added together across sources.
Approval status
A model name is listed for the inventory but never raises the "new tool" alert. The tool that called it is what you review. API keys and model deployments found through your own admin connections start as Approved. Use of an item you denied is recorded once for each person and day and raises a "Denied AI tool in use" alert. StackJack builds no score or ranking of people from this.
Work the list
- Open AI inventory → Inventory.
- Filter by Kind, Status and Where it came from. Tick Show apps that look unrelated to AI to see everything.
- Select a row to open its details.
- Select Approve or Deny. Denying an MCP server can block it on machines that receive your managed settings. If you turn down the last approved MCP server of an organization that allows approved servers only, the page asks you to type a confirmation phrase and press Deny again.
If settings are delivered to staff machines, the new lists go out with the next push. See Settings for staff machines.
What you will see
- A table with Tool, Status, Last 30 days, Risk class and Last seen.
- The "Waiting for a decision" number on the AI tools card of the Overview.
- The approved list in the policy you publish. See The AI use policy and acknowledgements.
Troubleshooting
Next
Was this helpful?
More in Control and respond
Rule packs, modes and custom rulesFindings and triageData-loss rules, findings and scanning StackJack tool trafficFilter personal data out of tool results (the PII filter)Still need help? Ask the team