AI Distilled
Playbooks · Make AI decisions with confidence
Make AI decisions with confidenceHuman approves

Write a one-page AI policy for your team

A one-page policy your team actually reads: what they may paste into AI, which tools are approved, and what always gets a human check.

Setup
30 min
Rung
01
Tool cost
Free
Approval
A person

Before and after

Before

  1. Everyone picks their own AI tools
  2. Nobody’s sure what’s OK to paste
  3. Questions get different answers
  4. Problems surface after the fact

After

  1. YouYou list your tools and data
  2. AIAI drafts a one-page policy
  3. YouYou make every call
  4. YouYou walk the team through it
  5. ToolA quarterly reminder to review it

Four steps, 30 minutes

  1. 01Take stock

    10 min

    Ask the team which AI tools they use and what for; a two-question message in team chat is enough. Then list the kinds of data you hold: customer details, staff records, financials, anything covered by a contract.

  2. 02Draft the policy

    5 min

    Give it your answers and ask for plain rules, not legal language.

    Help me write a one-page AI policy for [business], a [type of business] with [number] people.
Tools we use or want to use: [list the tools].
Data we handle: [list the kinds of data].
Write it in plain English, under 400 words, in four short sections:
1. Approved tools, and how to ask for a new one.
2. What you may paste, what you may paste only with names and details removed, and what you never paste.
3. What always gets a human check: anything sent to a customer; any name, number or date; anything about money, contracts or promises.
4. Who to ask, and what to do if something goes wrong.
Don’t cite laws or regulations. Where a rule might depend on one, write [CHECK].
    Open ChatGPTOpen ClaudeOpen Gemini

  3. 03Make the calls

    10 min

    The draft is a starting point. Decide every rule yourself, and check each approved tool’s data settings on the plan you actually use.

    Human check: Every rule is one you’ll actually follow and enforce, approved tools have had their data settings checked, and every [CHECK] is resolved. If you handle health, legal or financial data, or use AI in hiring, have someone qualified read it first.

  4. 04Share it and walk through it

    5 min

    Pin it in team chat and spend five minutes at your next team meeting going through the “never paste” list with examples from your own work. Date it, and set a reminder to review it each quarter.

What to measure beyond “time saved”

  • Questions asked

    Questions in the first month mean people are reading it. Silence can mean they haven’t.

  • Tool requests

    Requests to approve a tool show people are using the process, not working around it.

  • Near misses reported

    Someone stopping before pasting something private is a win. Make it easy to say so.

  • Updates made

    Tools and their data settings change. A policy untouched for six months is probably out of date.

Common pitfalls

  • Writing it for lawyers. Past one page, people skim. Keep to rules someone can follow on a busy day.

  • Banning everything. A blanket ban can push AI use out of sight, where nobody checks it. Approve a tool and set clear limits instead.

  • Setting it and forgetting it. Tools add features and change settings. Review it each quarter, and whenever you add a tool.

Need help building it?

AI Distilled can build this with you: a working pilot, tested on your real work.

Scope a build