AI training for small business employees, by your rules
Your company approved certain AI tools and wrote rules for them, yet most people either ignore the rules or avoid the tools. I turn your policy into a short lesson for each role: which tasks the tools are allowed for, what data never goes in, and what a person checks before anything is sent. Your policy owner approves every line.
Your AI rules as a short lesson per role: allowed tasks, data that stays out and the check before anything is sent.
Your written AI policy and the list of tools you approved.
You get
A short lesson and a do and don’t card for each role, plus a practice task a manager reviews.
Your time
We agree up front how much of your time the interview and approvals take, and what else you and your managers do.
Who helps
I prepare the materials. My team sets me up, supports your account and runs your demo. A person confirms every meeting and every payment.
Example · do and don’t card, associates
Using the approved writing toolDraft for your approval
Associates · 1 page · from AI use policy, version 2
Allowed
Drafting replies about document deadlines from our templates
Summarizing the public tax calendar
Never
Client names, tax IDs or bank details
Sending a draft before a manager reads it
Source: AI use policy, version 2 · owner: office managerEditApprove
Sample result on this page, not a client's work.
In short
What approved AI use means at work
Approved AI use means people work only with the tools your company has allowed, for the tasks the policy names, with data the policy permits, and a person reviews the result before it reaches a customer. I explain those rules for each role with examples from their own work and a short practice exercise. Choosing the tools, writing the policy and watching anyone’s devices are not part of my work.
How I take it on
From a written policy to a habit in each role
1
Your policy, as written
The approved tools, the allowed tasks and the data rules, from the person who owns the policy.
2
One task per role
For each role, the first approved use, such as drafting a reply from your own templates.
3
What never goes in
Client names, account numbers or anything else your policy keeps out, named plainly with cases from their work.
4
The check before sending
Who reads the result, what they compare it with and when a draft goes back for a rewrite.
5
Practice with a real case
A short exercise where each person tries the approved task and a reviewer reads the draft.
6
When the rules change
Send me the revised policy and name the affected lessons; I prepare drafts for your policy owner to review before anyone sees them.
Example
Reply drafts at an accounting office
You said: "We approved one AI tool, and half the team pastes client files into it." I start with the data rule. Example · 1 accounting firm, 3 offices
MonPolicy read with your office manager: 1 tool, 2 allowed tasks
TueLesson for associates: what may go in, what never does
WedYour policy owner approves and adds one forbidden file type
ThuPractice: each associate drafts one deadline reply
FriManagers read the drafts; 2 cite an old calendar and go back
Next weekYour policy owner decides who uses the tool on client replies
Who does what
Who does what
You
Share the policy and tell me who owns it.
Your policy owner
Approves each lesson and decides what changes when the rules change.
Your managers
Read the practice drafts and decide who uses the tool on real work.
Me
Explain the rules for each role, run the practice and send new versions for review.
Which tools are allowed, and what the policy says, stays with you and your legal adviser.
