Checklist

What to put in a staff AI use policy

A practical checklist for writing a staff AI use policy: scope, approved tools, data rules, human review, exceptions and review dates.

By the Cloudverg engineering team · · 4 min read

Who this is for

This is for the person who has been handed the job of writing the rules for AI at a small or mid-sized company. That is often an operations lead, an IT manager, a CTO or a finance lead, not a dedicated policy team. You need a document staff will read and follow, not one that sits in a folder.

What to put in a staff AI use policy is mostly about decisions you can make now: which tools are allowed, what data can go in, who checks the output and how people report a mistake. The policy works best when it is short, specific and tied to the tools your team already uses.

Before you begin

Start by listing the AI tools people already use, including the free ones they signed up for with a work email. Check expense reports, browser extensions and meeting notes tools. You cannot govern what you cannot see.

Then group your work by data type. Public marketing copy, internal meeting notes, customer records and source code all carry different risk. Write down which groups can use AI and which need a review first. Keep the list short enough to remember.

Decide who owns the policy. One named person should answer questions, approve exceptions and set the review date. A policy with no owner gets stale in a quarter. Pick a review cycle you can keep, such as every six months or after any major tool change.

The checklist

Use these items as the content checklist. You do not need every item on day one. Start with scope, approved tools and data rules, then add the rest before you widen access.

Write each item as a rule staff can act on. 'Use good judgment' is not a rule. 'Do not paste customer names, emails or account numbers into a public AI tool' is.

Keep the policy in one place and link it from the tools page, the onboarding checklist and the IT help desk. Staff should not have to hunt for it.

  • State the scope: which staff, contractors and tools the policy covers.
  • List approved AI tools by name and say what each tool is approved for.
  • Ban personal accounts for work data and explain how to request a new tool.
  • Define data tiers and name the data that may never go into a public model.
  • Require human review before AI output goes to a customer or shapes a decision.
  • Set disclosure rules for AI-generated text, images and code.
  • Require source checks for facts, quotes, citations and numbers.
  • Cover code use: license checks, secret scanning and no customer data in prompts.
  • Name the person who approves exceptions and the form or channel to use.
  • Set an incident path for wrong output, data leaks and tool outages.
  • Require a saved record or log for high-risk uses.
  • Add training and onboarding steps with a named contact.
  • Set a review date and name the owner who will run it.

What good looks like

A good policy fits on a few pages and uses examples from your own work. Staff can find the approved tool list in two clicks. The data rules use plain categories, not legal terms.

Exceptions get answered in a day or two. The owner keeps a short log of requests so the policy can improve. When a new tool appears, there is a clear path to test it, not a blanket ban that pushes use underground.

People know what to do when something goes wrong. They report it without fear of being blamed for trying to help. The policy also names the human who signs off on high-risk output, so accountability is clear.

What to do next

Draft the first version with the tool list, data rules and owner. Circulate it to a small group, including one person who uses AI daily and one who handles customer data. Fix the unclear parts, then publish.

Run a short briefing and add the policy to onboarding. Review the exception log after the first month to see which rules cause friction. Update the policy and tell people what changed.

If you want help ranking use cases and planning a first project, the AI readiness assessment from Cloudverg, at $5,500/month, is one option alongside an internal workshop or a consultant. It runs two weeks and covers data readiness, a reference architecture and a cost model.

Frequently asked questions

What should a staff AI use policy include?

It should name approved tools, set data rules, require human review for customer-facing output, explain how to request an exception and name an owner and review date. Keep it short and tie it to the tools people actually use.

Should we ban AI tools at work?

A blanket ban is hard to enforce and pushes use to personal accounts. A better approach is to approve a small set of tools, restrict sensitive data and give staff a clear path to ask for something new.

How often should we review the policy?

Review it at least twice a year and after any major tool change or incident. Set a named owner and a calendar reminder so the review actually happens.

Monthly plan

Want help with this? AI readiness assessment

$5,500/month

2 weeks · Starts within 10 business days

Two weeks to rank your AI and agent ideas by value and risk, check your data, and plan a first project with a cost model.

Renews monthly. Cancel any time.
What is included

Related guides

More ai and agents guides   AI and machine learning

Smallest way to start

Get a senior architect on your problem this week

Ninety minutes a month on the cloud, AI or security question that is blocking you, a pre-read of your material and a written summary within two business days.

$450/month

Renews monthly. Cancel any time.