What can non-fee-earning staff in a law firm use AI for?

Non-fee-earning staff can use AI for most of the preparation work they already do: pulling facts out of a file, building a chronology, drafting routine correspondence, organising bundles, summarising what is in a box of papers, and preparing internal reports. What does not change is that anything touching a client's matter still needs a fee earner's judgement and sign-off, and that the firm's duties on confidentiality and supervision apply to every person with access, not just to the qualified ones.

The practical question is not whether support staff should be allowed near AI. In most firms they are the people doing the document-heavy tasks that AI helps with. The question is what the output is allowed to be used for once it exists.

The work that suits support staff best

Most of the useful ground is preparation rather than opinion:

• Finding things in a file. Where is the signed engagement letter, which letter chased the other side in March, what did the client say about the deposit. A legal assistant can get an answer with a page reference faster than scrolling.
• Chronologies and schedules. Building a dated list of events from correspondence is slow, mechanical and exactly the sort of thing that can be drafted and then checked.
• Bundle and disclosure preparation. Identifying duplicates, spotting gaps in a run of documents, listing what is in a folder.
• Routine correspondence. Acknowledgements, chasers, appointment confirmations, requests for missing documents.
• Internal reporting. A practice manager can use AI to help draft a file review template, a procedure note, an induction pack or a summary of what a long policy document requires.
• Getting a handle on a new file. When a matter arrives from another firm, a paralegal can read in far faster with an assistant that cites pages.

Notice what these have in common. They are all tasks where the answer is checkable against the papers, and where a mistake is visible to anyone who looks.

The line that matters

The line is not fee earner versus non-fee earner. It is preparation versus advice and decision.

A paralegal may use AI to draft a letter. A fee earner decides whether that letter should be sent, and in what terms. A legal assistant may use AI to extract the dates from a file. A fee earner decides what the limitation position is and what goes in the diary. A practice manager may use AI to summarise a long piece of official guidance. A compliance officer decides what the firm does about it.

Put plainly in your policy: AI output produced by support staff is internal working material until a fee earner adopts it. Nothing goes to a client, the other side or the court on the strength of an AI draft alone.

The second line worth drawing is about answering questions that look legal. If a client telephones and asks whether they can still bring a claim, that is not a question for a support member of staff to put to an AI tool and relay. It goes to the fee earner. Tools make it easy to produce a confident-sounding answer, which is precisely why the escalation rule needs to be explicit.

Confidentiality and access do not relax

Support staff often have the broadest access in a small firm, because they handle post, scanning, billing and file opening. That makes the access controls in any AI tool more important, not less.

Ask, before you roll anything out:

1. Can a person see only the matters they are working on, or does everyone with a login see everything?
2. Does seniority on its own give a wider view, and do you want it to?
3. Is every account tied to a verified email address, so a shared login is not quietly in circulation?
4. Where does the information go, and is it kept apart from other firms' work?
5. What happens to access when a temp, an agency paralegal or a work experience placement leaves?

If your conflict and information barrier arrangements mean a particular assistant should not see a particular file, the AI tool has to respect that. A tool that cannot be configured that way is a tool that undermines your barriers.

Supervision looks slightly different here

Supervising a fee earner's use of AI mostly means checking their judgement. Supervising support staff means checking the handover: did the right material go in, and is the output being passed up rather than acted on.

Three habits are worth building in:

• Say where the answer came from. If a chronology or a summary is produced with AI help, the fee earner receiving it should be able to see the page each point rests on and spot check a handful.
• Flag what is missing. An honest "the papers do not cover this" from an assistant is more useful than a tidy document with a quiet gap in it.
• Keep one note of the task. A short line on the file saying what was prepared, by whom, with what tool, and who checked it. That is enough for a file review to make sense a year later.

Firms sometimes worry that giving support staff these tools will blur roles. In practice the opposite tends to happen. When the mechanical work is done faster and comes up with page references attached, the fee earner spends more of their time on the part that actually requires them.

Where Alesis fits

Alesis is an AI assistant for UK law firms, used through the web browser, with one conversation for a matter, so a paralegal, an assistant and a fee earner can work from the same thread on the same file. Inside a firm people see only the matters they are on, and seniority alone grants no view. It answers questions about a matter from the matter's own papers and names the page each answer came from, and where the papers do not say, it says so. It prepares drafts for a qualified person to review and sign off, and it does not file, serve or email anything.