AI Client Intake for Florida Law Firms: A Practical Guide to Bar Opinion 24-1

A prospective client usually reaches out when something has just gone wrong: a car accident, a divorce filing, or a contract dispute. If nobody answers until the next morning, they may call another firm. That's why many small firms in Jacksonville and Orlando are looking at AI chat or automated intake for their websites and phone lines.

Intake is also where the ethical risk is highest. In January 2024, The Florida Bar issued Ethics Opinion 24-1 on lawyers' use of generative AI. The opinion says lawyers may use AI, but they must protect client confidentiality, supervise the tool, avoid improper billing, and follow the advertising rules. It specifically addresses chatbots used for client intake.

This guide explains how to set up AI intake for a small firm in a way that is useful and that you can defend. It isn't legal advice, and you remain the expert on your own ethical duties. Our focus is the systems side: the workflow, the data, and the controls.

What Opinion 24-1 says about intake chatbots

The opinion is advisory, not binding, but it is the clearest guidance Florida lawyers have. A few points apply directly to intake:

• Say it's AI. Chatbots that talk with clients or third parties must make clear that they are an AI program and not a lawyer or employee of the firm.
• Collect facts, not give advice. The opinion points to Florida Ethics Opinion 88-6, which allows nonlawyers to conduct an initial interview if they identify themselves as nonlawyers, limit questions to factual information, and refer legal questions to a lawyer. The same limits make sense for a chatbot.
• Watch for accidental relationships. The opinion warns that an "overly welcoming" chatbot could create a prospective-client or even a lawyer-client relationship without the lawyer knowing. Clear disclaimers that limit the firm's obligations matter.
• Screen for existing representation. The opinion suggests screening questions that limit the chatbot's communication if the person already has a lawyer.
• Know where the data goes. Lawyers should research the tool's data retention, data sharing, and "self-learning" policies, and the opinion recommends informed consent before using a third-party AI program that would receive confidential information.

Those points translate neatly into a design checklist.

Start with one intake workflow

Don't automate everything at once. Pick one entry point and measure it. For most small firms, the best candidates are:

1. After-hours website inquiries: visitors who fill out a form or open a chat outside office hours.
2. Missed calls: callers who hang up or reach voicemail.
3. Practice-area triage: routing an inquiry to the right attorney or declining matters the firm doesn't handle.

Choose the one that loses the most leads today. If you don't know which that is, spend two weeks counting first. Our article on setting a two-week baseline explains how to do that with the tools you already have.

Designing a compliant AI intake flow

1. Disclose up front, every time

The first message should say clearly that the visitor is talking with an AI assistant, not a lawyer or staff member, that it can't give legal advice, and that using it doesn't create an attorney-client relationship. Here's an example opening message (hypothetical):

"Hi, I'm an automated AI assistant for the firm. I'm not a lawyer or a firm employee, and I can't give legal advice. I can collect some basic information so an attorney can decide whether we're able to help. Sharing information here doesn't make you a client of the firm. Please don't share details you consider confidential until we confirm we can take your matter."

Have an attorney approve the final wording. The point is that the disclaimer comes before the visitor types anything substantive, not buried in a footer.

2. Ask only for what a conflict check and callback need

A good intake flow collects the minimum: name, contact information, the type of matter, the names of other parties (for a conflict check), key dates such as a court deadline, and whether the person already has a lawyer. If the answer to that last question is yes, the assistant should stop and politely refer them back to their attorney.

Collecting less protects both sides. The less sensitive detail sits in a chat log, the smaller the exposure if something goes wrong.

3. Block legal advice by design

Configure the assistant to decline questions like "Do I have a case?" or "What should I do about this notice?" and route them to a lawyer. Test this deliberately. Try twenty or thirty realistic questions that invite advice and confirm the assistant declines every one. Repeat the test whenever the underlying model or prompt changes.

4. Flag urgency for a human

Some inquiries can't wait until morning: an upcoming hearing, an arrest, or a protective order. Define those triggers and have the system alert an on-call attorney or staff member rather than letting the assistant handle them.

5. Put a lawyer in the loop before anything goes out

The assistant gathers facts. A person reviews them, runs the conflict check, and decides whether to call back. Opinion 24-1 compares AI to a nonlawyer assistant whose work the lawyer must supervise, and that's the right mental model.

If the firm follows up by text message, keep it short, factual, and give people an easy way to stop. Here's an example (hypothetical):

"This is [Firm Name]. We received your inquiry and an attorney will review it today. Please don't send case details by text. Reply STOP to opt out."

The security questions to answer before you buy

Opinion 24-1 asks lawyers to understand the technology well enough to meet their duties. In practice, that means getting written answers to these questions from any vendor:

• Retention: How long are chat transcripts and recordings kept, and can you set a shorter period or delete them on request?
• Training: Is your data used to train or improve the vendor's models? Can that be turned off by contract, not just a setting?
• Access: Who at the vendor can see your conversations, and is access logged?
• Breach notice: Will the vendor notify you of a breach or of a subpoena for your data? The opinion carries over guidance from Florida Ethics Opinion 12-3 on cloud providers, including enforceable confidentiality obligations and notice of breaches.
• Data location and exit: Where is data stored, and what happens to it if you cancel?
• Integration: Does intake data flow directly into your practice management system, or does it pile up in another inbox?

On your side, turn on multi-factor authentication for every account that can see intake data, limit access to the people who need it, and decide who reviews new inquiries each day.

How to measure whether it worked

Before launch, record your baseline: how many inquiries arrive after hours, how many get a response within one business day, and how many become consultations. After launch, track the same numbers for the same length of time, along with two quality checks: how often the assistant tried to give advice (the goal is zero) and how often a human had to correct its summary.

If response time improves and the quality checks stay clean, expand to the next entry point. If not, you've learned something useful without betting the whole intake process on it.

How R Software can help

We help small businesses in Jacksonville, Orlando, and across Florida start with one workflow, measure the result, and build security in from the beginning. For a law firm, that often means mapping your current intake path, reviewing vendor data practices against the questions above, and setting up an intake flow your attorneys can supervise.

Want a second opinion on where to start? Take our free assessment at rsoftware.net/assessment or email us at hello@rsoftware.net.

For more practical guides on secure AI and automation for local businesses, follow R Software & Consulting on LinkedIn.

FAQ

Can a Florida law firm use an AI chatbot for client intake?

Yes, according to Florida Bar Ethics Opinion 24-1, as long as the firm protects confidentiality, supervises the tool, and follows the advertising rules. The chatbot must identify itself as AI and not a lawyer or employee, and it shouldn't give legal advice.

Do we need client consent to use AI?

The opinion recommends getting a client's informed consent before using a third-party generative AI program if doing so would disclose confidential information. It notes that consent isn't required under Rule 4-1.6 when no confidential information is disclosed to a third party. Talk with your ethics counsel about your specific setup.

Is a self-hosted AI tool safer than a cloud service?

The opinion notes that confidentiality concerns may be reduced with an in-house tool instead of one where a third party hosts and stores the data. Self-hosting brings its own security responsibilities, though, so the right choice depends on your firm's size and IT support.

What's the first step?

Count your current after-hours and missed-call inquiries for two weeks. That tells you whether intake automation is worth it and gives you a baseline to measure against.

Written with help from our AI assistant.