Blog / 8 min read
What Should an AI Policy for a Small Law Firm Actually Say?
By Shang Ahmed · 9 November 2026
What a small law firm's AI policy needs to cover, in plain terms: approved tools, client data, review, consent, training and ownership.
Quick Answer
A small law firm's AI policy should answer eight questions in plain language: which AI tools are approved, what information may go into each one, how output is checked before it leaves the firm, when clients are told, who can approve new tools, how staff are trained, what to do if something goes wrong and who owns the policy. It does not need to be long. Two pages that everyone actually follows are worth more than twenty that nobody reads.
What Is an AI Policy?
An AI policy is a short internal document that sets out which AI tools a firm's people may use, for what purposes and under what rules. For a law firm, its main job is to protect client confidentiality and the quality of legal work while still letting lawyers benefit from the time AI can save.
Why Does a Small Firm Need One?
A small firm needs an AI policy because its people are almost certainly using AI already, and without a policy each person is making their own decisions about client information. Regulators now expect firms to have controls in place, whatever their size. Most firms still do not have a written AI strategy.
The numbers show how wide the gap is. The 8am 2026 Legal Industry Report found that 69% of legal professionals use generative AI for work, but only 9% say their firm has a written AI policy that is actively enforced. 43% said their firm has no policy and no plans for one.
Regulators are moving the other way. The SRA's warning notice on the misuse of AI, published in August 2026, reminds firms that they need effective governance structures, systems and controls to manage AI risks. In the EU, Article 4 of the AI Act requires organisations that use AI to take measures to support AI literacy among their staff. See what Article 4 means for law firms.
What Should the Policy Say?
The policy should cover eight areas, each answered in a few sentences. Write it for the person who will read it on a busy afternoon, not for a regulator.
1. Which tools are approved?
List approved tools by name, including AI features inside software the firm already uses, such as email, document and meeting tools. State clearly that anything not on the list needs approval before it is used for work.
2. What information may go into each tool?
Sort information into simple categories, for example:
- Public information, such as published judgments or legislation
- Internal firm information, such as templates and know-how
- Confidential client information
Then say which categories each approved tool may handle. Most firms allow only tools under a proper business contract to handle client information, and never consumer chatbots used with personal accounts.
3. How is output checked?
State that a lawyer must review all AI assisted work before it leaves the firm, and that every citation, quotation and factual claim must be checked against the original source. Courts have taken a hard line on fake citations, including the High Court in London in Ayinde v London Borough of Haringey [2025] EWHC 1383 (Admin).
4. When are clients told?
Explain when the firm will tell clients about AI use and when it will ask for consent. A simple rule is to be open with clients whenever confidential information goes into an AI tool, and to ask for consent where the firm does not fully control the tool.
5. Who can approve a new tool?
Name the person or role who decides. Set out the minimum checks before approval: the contract and data processing terms, whether inputs are used for training, how long data is kept, where it is processed and who can access it. The same checks sit behind the questions to ask before adopting any legal AI tool.
6. How are people trained?
Commit to a short practical session for everyone, repeated when tools change, and keep a record of who attended. This also supports the AI literacy expectation under the EU AI Act.
7. What happens if something goes wrong?
Tell staff exactly what to do if client information goes into the wrong tool or an AI error reaches a client: report it to a named person straight away, so the firm can assess any confidentiality, data protection or reporting obligations. Make clear that early reporting is expected and supported.
8. Who owns the policy?
Name an owner and a review date. A quarterly review is sensible, because AI tools and their terms change often.
What Should the Policy Not Do?
The policy should not ban AI outright, copy a large firm's template or read like a legal treatise. Each of these makes it less likely that people will follow it.
- Outright bans tend to push people towards tools on their personal phones, which is worse than a controlled approval process.
- Large firm templates assume a compliance team and IT department that a small firm does not have.
- Legalistic language means staff skim it and miss the parts that matter.
How Do You Write and Roll Out the Policy?
Start by finding out what people actually use, write the first version in an afternoon, walk everyone through it and review it after three months. A short, real policy beats a perfect one that never gets finished.
- Ask everyone which AI tools and features they use, including support staff.
- Check the terms of the tools in use and decide which to approve, restrict or stop.
- Draft the eight sections above in plain language, ideally on two pages.
- Walk through it in a 30 minute meeting using real examples from your own work.
- Make the approved tool list easy to find, for example pinned in your shared drive.
- Record who has read it and when.
- Review it every quarter or when a key tool changes.
FAQ
How long should an AI policy be?
For a small firm, one or two pages is usually enough. The test is whether a new trainee could follow it without asking questions.
Do we need a lawyer to write our AI policy?
Not necessarily. The policy is mainly operational. But the parts on confidentiality, data protection and client consent should reflect the professional and legal rules in your jurisdiction.
Should the policy name specific tools?
Yes. A list of approved tools by name is far easier to follow than general principles.
How often should we update it?
At least every quarter, and whenever the firm adopts a new tool or a provider changes its terms.
Related reading
The Bottom Line
An AI policy for a small law firm does not need to be long or complicated. It needs to answer eight practical questions so that every person in the firm knows which tools they may use, what may go into them and who checks the output. Most firms can write a solid first version in an afternoon.
Want a clear starting point for your firm's AI policy based on the tools your people actually use? Our AI Tools Assessment maps your tools and workflows and gives you a practical plan.