The short answer
A small company needs a simple, one-page AI policy before training staff, covering five things: what data must never go into public AI tools, which tools are approved, that all AI output is checked by a person before use, who is accountable, and where to ask questions. It should be plain English, not a legal document. The aim is clear boundaries people actually remember, set before wide use begins.
Why a small company needs an AI policy before training
You do not need a forty-page governance framework. You are a small company, not a bank. But you do need something, and you need it before you train the team and open the floodgates, because a policy written after an incident is a policy written too late.
The purpose is not bureaucracy. It is to give people clear, memorable boundaries so they can use AI confidently without accidentally causing a problem. A good small-company AI policy fits on a page and can be read in two minutes.
What your one-page AI policy should cover
Five things, in plain English:-
- What never goes in. The data that must not be pasted into public AI tools: client and customer personal data, confidential figures, contracts, passwords, anything you would not put on a public website. This is the most important line.
- Which tools are approved. Name the one or two tools the company uses, and any paid or business versions preferred for sensitive work. It stops people scattering data across random apps.
- A human checks everything. State plainly that AI output is always a first draft, checked and approved by a person before it goes to a client or into a decision. Nobody files AI output unread.
- Who is accountable. Make clear the person using AI owns the result. Their name is on the work, so the responsibility to check it is theirs. This keeps standards honest.
- Where to ask. Name a person or channel for questions, so an unsure member of staff asks rather than guesses. Questions are cheap; mistakes are not.
Keep it to a page and plain English. A policy nobody reads protects nobody. The best AI policy is the one your team can actually remember on a busy Tuesday.
Policy and training go together
A policy on its own is a document. Training on its own can leave gaps. Together they work: the policy sets the boundaries, and the training shows people how to work within them while getting real value. That is why we handle both, and why the policy should be in place before the training so the two reinforce each other from day one.
The Oxford AI School helps small companies write a simple, sensible AI policy and then trains the team to use AI safely and well inside it. If you want the guardrails right before you get everyone going, book a call and we will help you set both.
What AI policy should a small company have before training staff? - frequently asked questions
What AI policy does a small company need before training staff?
A simple one-page policy covering what data must never go into public tools, which tools are approved, that a human checks all output, who is accountable, and where to ask questions. Plain English, not a legal document.
Do small businesses really need an AI policy?
Yes, a simple one. It gives staff clear boundaries so they can use AI confidently without accidentally leaking data or sending unchecked output. It should be in place before wide use, not written after a problem.
What should an AI usage policy include?
Prohibited data, approved tools, a human-checking rule, accountability, and a place to ask questions. Those five points, kept to a page, cover the real risks for a small company.
How long should a small business AI policy be?
One page. A short, plain-English policy that people actually read and remember protects you far better than a long governance document nobody opens.
Should the AI policy come before or after training?
Before. The policy sets the boundaries and the training teaches people to work within them. Having the policy first means the team starts safely from day one rather than developing habits you later unpick.
Which Oxford AI School training fits this need?
Choose practical training around your team, tools and real work. Sessions are available in Oxfordshire, London, across the UK, in the United States and internationally through live online delivery.
Get your guardrails right first
Book a call and we will help you write a simple AI policy and train your team to work safely within it.
Book A Call
Harry Lang runs The Oxford AI School, teaching teams and businesses the basics of AI in a fun, easy and practical way. A marketing leader turned AI trainer, he can be reached at Learn@TheOxfordAISchool.com.