IVIS

Free checklist

The pre-launch AI checklist.

A practical, plain-English checklist to run before you ship something built with AI. Use it as a self-check — or have ZIVIS run a full Mini AI Risk Map for you.

What it touches

  • List the data the system reads, writes, stores, and sends.
  • Note any customer, financial, HR, or otherwise sensitive data in scope.
  • Map the tools, APIs, and systems it connects to.
  • Identify the credentials it uses and their access scope.

Who can use it

  • List the users and roles who can trigger it.
  • Confirm whether external parties can reach it.
  • Check that permissions match what the workflow actually needs.
  • Identify the trust boundaries — where untrusted input enters.

What could go wrong

  • Can the AI be steered by instructions hidden in content or data?
  • Could it reveal data it shouldn't?
  • Can it take a state-changing action without human approval?
  • What happens if it makes the wrong decision — and who notices?

Before you ship

  • Scope credentials down to least privilege.
  • Add an approval step for outbound or state-changing actions.
  • Log prompts, tool calls, and decisions so incidents are reconstructable.
  • Decide whether the risk warrants a deeper audit or AI testing.

Want a second set of eyes?

The Mini AI Risk Map turns this checklist into a real, prioritized map of your system for $1,499.