Talking to Clients About Legal AI: Transparency Without Alarm

As soon as you start using AI on client work, a new question appears:

“What do we tell clients about it?”

Say nothing and you risk clients feeling misled if they find out later. Say too much, too dramatically, and you risk alarming people or suggesting that robots are writing their advice.

This article offers a practical approach to talking to clients about legal AI — clear, calm, and aligned with UK professional obligations.

1. Be honest about what you are doing

A good starting point is a simple, truthful description along the lines of:

“We sometimes use AI tools to help us summarise documents, draft first-pass notes or keep track of tasks, but qualified lawyers always remain responsible for the advice and documents you receive.”

This makes three things clear:

AI is assisting, not replacing, human lawyers;
its use is limited to certain types of tasks;
responsibility for outcomes remains with the firm.

You can then expand or tailor this description depending on the client’s level of interest and sophistication.

2. Focus on benefits clients care about

Most clients do not want a lecture on large language models. They care about:

speed — “Will you move faster?”
cost — “Will this cost me more or less?”
quality — “Will this make mistakes?”

Helpful ways to frame benefits include:

“We use AI tools to reduce time spent on mechanical tasks like summarising long email chains or bundling documents, so more of your budget goes on analysis and strategy.”
“AI helps us keep your matter record and chronology neat and up to date, which in turn supports better decisions.”

Avoid vague marketing phrases (“we are an AI-driven firm”). Stick to specific, client-facing benefits.

3. Acknowledge limitations and safeguards

Clients will have heard about AI “hallucinations”. They may worry that:

you will rely blindly on machine outputs; or
their information will be sprayed around the internet.

Address this head-on by explaining, briefly:

that AI outputs are checked by lawyers before being relied upon;
that you use tools within governed environments (for example, inside your case management system or Microsoft 365), not consumer chatbots;
that you do not use AI to replace legal judgment or to draft advice without review.

A simple statement such as:

“We treat AI suggestions in the same way as we treat a junior’s draft: helpful, but always checked.”

goes a long way.

4. Be clear about confidentiality and data handling

For in-house legal teams and institutional clients, confidentiality is often the main concern.

Without going into vendor names unless necessary, you should be able to say:

where AI processing happens in broad terms (for example, “within UK or EU data centres under our contracts with providers”);
that you do not use tools which train on or share clients’ confidential material with other customers;
that prompts and outputs are treated as part of the client’s file, with appropriate access controls.

Some clients will want more detail in writing, perhaps as part of outside counsel guidelines. Having a short, non-technical AI use statement ready will help.

5. Tailor the message to different client types

Not all clients need the same level of explanation.

Retail and high-street clients may only need a short reassurance that AI is used “behind the scenes” to make the work more efficient, and that lawyers remain responsible.
Institutional and repeat clients are more likely to ask about data protection, model choice and audit trails. They may have their own policies you need to align with.
Tech-savvy clients might be positively interested and ask how you are using AI — here, you can use more detail to show you are being thoughtful and responsible.

The core message — assistance, not replacement; confidentiality preserved; human responsibility — remains the same.

6. Decide what goes into terms and what stays conversational

You may choose to:

add a short paragraph to engagement letters or terms of business explaining that you may use AI tools as part of your service delivery, with appropriate safeguards;
maintain a more detailed internal policy and a client-facing summary on your website.

Avoid:

over-complicated disclaimers that sound as if you are trying to offload responsibility onto clients;
vague “we might use AI” clauses that do not inspire trust.

Think of your AI wording as an extension of how you already describe outsourcing, cloud services and document management.

7. Train your lawyers to talk about AI consistently

Even if you have good written wording, individual fee-earners may improvise under pressure. To keep things consistent:

give people short scripts and examples they can adapt;
run short training sessions on how to answer common client questions about AI;
make sure everyone understands what tools are actually in use and what guardrails are in place.

The aim is that any partner or associate can explain, in a sentence or two, how the firm uses AI — without over-claiming or underselling.

Where OrdoLux fits

OrdoLux is being designed to make these conversations easier because:

AI features run inside the matter file, not in opaque external tools;
prompts, outputs and supervision notes are saved alongside other matter records;
logs show how AI was used on a file, making it easier to answer client queries or audits;
you can align actual workflows with whatever you tell clients about AI use.

So when you say “we use AI to help with summaries and chronologies, but lawyers remain responsible and everything sits in our core systems”, that is genuinely how OrdoLux is set up to work.

This article is general information for practitioners — not legal advice, regulatory advice or PR consultancy for any particular firm–client relationship.

Looking for legal case management software?

OrdoLux is legal case management software for UK solicitors, designed to make matter management, documents, time recording and AI assistance feel like one joined-up system. Learn more on the OrdoLux website.