AI for small law firms: where to start and what to skip

A small law firm should start with AI where it loses the most work for the least reason: unanswered calls, slow replies to new enquiries and fee-earners interrupted for updates. Those uses pay back quickly and are easy to check. Skip expensive document platforms until you have a specific drafting or review problem, and never let any tool advise a client unchecked.

Small firms are in an odd position with AI. The SRA's Risk Outlook on AI noted that dedicated legal AI was once accessible only to the largest firms, and that this is no longer the case. The tools are within reach. What small firms lack is time to evaluate them, a partner free to own the decision, and room for a failed six-month project. This guide is written for that reality. For the wider picture, see our guide to AI for UK law firms.

Is AI worth it for a small law firm?

It is worth it where the firm is losing something it can count. For most small firms that is not drafting speed. It is the enquiry that rang at 5.40pm, reached voicemail and instructed somebody else by the morning. It is the conveyancing client who rang four times for an update and left a review about it. It is the sole practitioner who cannot take a call in court and cannot afford a full-time receptionist to take it for them.

A small firm has one structural advantage here: the partners can see every part of the business. A two-week count of missed calls, a look at how long new enquiries wait for a reply, and a conversation with whoever picks up the phone will tell you more about where AI would help than any vendor demo.

What should a small law firm use AI for first?

We suggest ranking candidate uses on three things, each scored 1 to 3: how often the task happens, how much a mistake would cost, and how easily a person can check the output. Frequent, cheap-to-get-wrong and easy-to-check wins. In most small firms, the same three uses come out on top.

• The front door. An AI that answers the firm's line when nobody can, says it is AI, asks your qualifying questions and books the call back. The output is a summary a fee-earner reads, so checking is built in. We look at what such a system should and should not say in AI chatbots for law firms.
• Fee-earner cover and client updates. A sole practitioner or small team spends a surprising share of the week answering 'what is happening with my matter'. Covering those calls with a set of approved messages frees real hours.
• Routine correspondence. A general assistant on a business account, used for first drafts that a person always edits.

What these have in common is that they sit around the legal work rather than inside it. That is deliberate. A small firm has no spare capacity to absorb a mistake in the legal work itself, and no second partner to catch it.

What AI should a small law firm skip for now?

• Enterprise document platforms. They are built for large teams doing large volumes of review and drafting, and they are typically sold on annual contracts. Harvey, for example, does not publish its pricing. If you do not have a named drafting or review problem, you do not need one yet. We explain the category in what Harvey AI is.
• Anything that produces legal authorities without you checking them. The Ayinde judgment made the cost of that very clear, and a small firm has the least capacity to absorb a referral or a wasted-costs order.
• Autonomous tools that act without approval. Sending, filing or committing the firm to anything without a person pressing the button is a risk a small firm does not need. See AI agents for law firms for how to tell the difference.
• Five pilots at once. You will not have time to review any of them properly.

How much does AI cost a small law firm?

It varies more than any other software line. General assistants are sold per user per month at consumer or business prices published on the vendors' sites. Legal document platforms usually do not publish prices and are negotiated. Front-door and call-handling tools tend to be priced per firm, per fee-earner or by usage. Compare on the whole-year cost at your likely volume, not the headline figure.

A useful way to judge any price is to ask what it has to recover. A worked example, with invented numbers for illustration only: a hypothetical four-solicitor high street firm estimates it misses 25 calls a month from new callers outside the hours someone can answer, and that one in ten of those would have become an instruction worth, on average, £1,500 in fees. That is roughly £3,750 a month of work walking to other firms on the firm's own assumptions. Your numbers will differ. The point is to write them down before the sales call rather than after. Our piece on what missed enquiries cost a law firm goes into how to estimate this properly.

What are the risks of AI for a small firm?

The same as for any firm, with less slack. The SRA has been clear that the named solicitors in a firm remain accountable for what its systems produce, and a small firm's COLP is often also its senior partner, its supervisor and its IT department.

• Staff using free tools on personal accounts with client information. Ask, without blame, and give them an approved alternative.
• A client-facing tool that does not say it is AI, or that can be talked into advice.
• No record of what the tool said. If you cannot see what a caller was told, you cannot defend it.
• Supplier lock-in. Check you can export your data and leave.

None of these requires a large compliance project. They require a one-page policy and a named owner. AI governance for law firms has a template short enough to adopt in an afternoon.

How to run a one-month AI trial in a small firm

This is the method we would follow if we ran a small firm. It assumes nobody has more than an hour a week to give it.

• Before you start, measure one thing for two weeks: missed calls, time to first reply for new enquiries, or update calls per fee-earner. Write the number down.
• Choose one tool for that one problem. Agree in writing what it is allowed to do and what it must hand to a person.
• Name one reviewer. For the first fortnight they read every output (every call summary, every draft). After that, a sample.
• Keep a short log of anything that went wrong or nearly did, and what the tool did at the edge of its instructions.
• At the end of the month, compare against your baseline and decide: extend, change or stop. Write down why.

The trap most small firms fall into is judging the trial on how impressive the tool felt. Judge it on the number you wrote down at the start. A tool that felt ordinary but halved your unanswered enquiries has done more than one that wrote beautiful paragraphs nobody needed. If the number did not move, stop the trial, keep the notes, and try the next problem on your list with what you learned.

What should a small firm tell its staff about AI?

More than most partners expect. In a small firm the person answering the phone, the paralegal and the trainee are often the heaviest users of general AI tools, and they are the least likely to have been told what is allowed. The SRA's Risk Outlook made the same observation: staff may use freely available tools occasionally and casually, even when the firm has not formally adopted AI. Silence from the partners does not mean nobody is using it. It means nobody is checking.

A short conversation covers most of it. Tell staff which tools are approved and on which accounts. Tell them what never goes into any tool: client names with matter detail, anything a client told the firm in confidence, identity documents. Tell them that everything a tool produces is a draft until a person has read it. And tell them who to ask when they are unsure, because the question they do not ask is the one that becomes a problem.

If you use a client-facing tool, such as an AI answering the phone out of hours, make sure the whole team knows what it says to callers and what it hands over. The worst outcome is a client who spoke to the firm's assistant last night ringing this morning and finding that nobody knows the call happened. The summary should land where the team will see it first thing.

Can a sole practitioner use AI?

Yes, and a sole practitioner often benefits most because there is nobody else to take the call. The obligations do not change with size: disclose that a caller is speaking to AI, keep advice with the solicitor, keep client information under your control, and check anything that goes out. What changes is how much a single missed call or a single bad week costs, which is exactly why cover for the phone tends to come first.

Where WasilX fits for a small firm

We built for this shape of firm. Cam takes calls on your line at any hour, says it is AI, asks your qualifying questions, books the call back and sends you a summary. A personal assistant covers each fee-earner's calls while they are in court and drafts matter updates for the fee-earner to approve. We do not draft legal documents or give legal answers. If the phone is where your firm loses work, we are happy to talk it through.

Frequently asked questions

What is the best AI for a small law firm? There is no single best tool. The best first tool is the one that fixes the problem you can count, which for most small firms is unanswered calls or slow replies to new enquiries rather than drafting.

Do small law firms need an AI policy? Yes, but a short one. A page that lists approved tools, what they may be used for, who reviews their output and who owns the decision covers most of what a small firm needs.

Does a small law firm still need a receptionist if it uses AI? Usually, yes. AI can cover the hours and the calls a person cannot, such as evenings, lunch and busy mornings. It is meant to run alongside staff, not in place of them.

Is free AI safe for law firms to use? Free consumer accounts may not give the firm control over how data is kept or used. Check the terms, and do not put client information into any tool the firm has not approved.