A law firm chatbot looks like a website feature. Under the advertising rules it is closer to a billboard that talks back. The Florida Bar took this up in Ethics Opinion 24-1. A chatbot that talks with clients or third parties has to follow the advertising rules. It also has to say that it is an AI program, not a lawyer or a firm employee. Advisory opinions bind nobody. Still, this one is the clearest map on offer.
A Law Firm Chatbot Is Advertising, Not Software
Marketing vendors sell chat as a conversion tool. Bar regulators see something else. They see a message from the firm to a member of the public, and that is what the advertising rules cover.
So the duty does not sit with the vendor. The vendor wrote the code. The firm makes the statements. Nothing in the rules shifts that because a third party runs the tool.
That puts the chat window at the same height as every other firm message. Nothing misleading. No claim about the firm’s skill or record unless you can back it up.
It Has to Say What It Is
The Florida opinion is direct on this point. A chatbot speaking for the firm has to carry a disclaimer. It is an AI program, and it is not a lawyer or an employee.
Most installed widgets fail that on day one. They open with a friendly first name and a smiling avatar. That is the exact impression the rule warns about.
The fix costs nothing. Say it in the first message. Keep it on screen rather than behind a link. Drop the human photo. Visitors do not mind talking to a bot. They mind finding out later.
It Can Create a Prospective Client Without You
Here is the risk that catches firms out, and it has little to do with technology. It comes from the rule on duties to prospective clients.
The comment to that rule explains when a consultation happens. A firm invites someone to send details about a possible matter. The person sends them. No clear warning limits what the firm takes on. That is enough.
Now look at a chat window. It is an invitation, and a stranger typing the facts of their case has answered it. Whether a relationship exists has long turned on what that person reasonably believed. So a warm bot that sounds like a lawyer makes the belief easier to hold.
Screening Questions Belong at the Top
The opinion points to a simple safeguard. Use screening questions that cut the conversation short when another lawyer already represents the person.
That is worth doing for plain business reasons too. Conflicts surface sooner. Your intake team stops burning mornings on matters the firm was never going to take.
Set the handoff rule at the same time. Decide which questions a human has to answer, and have the bot say that a person will follow up. A bot that knows its limits loses fewer people than one that guesses.
- State in the first line that the visitor is chatting with an AI program.
- Warn, before any facts arrive, that the exchange is not confidential.
- Ask early whether another lawyer already represents the person.
- Collect contact details and a short description, then stop.
- Hand off to a human for anything that looks like legal advice.
Claims About Your AI Have to Be Verifiable
A firm may advertise that it uses AI. A firm may not claim its AI beats the tools other firms use, unless it can prove the claim.
So drop the superlatives. A phrase like the most advanced legal AI has no place on a firm website. Nobody can back it up on request.
Describe what the tool does instead. It answers common questions. Then it takes details before a call. And it books consultations after hours. Plain descriptions read better than adjectives anyway.
That same standard covers the rest of your marketing copy. Words about results, reputation, or skill need something behind them. The rule is older than any chatbot. All the bot did was move it to a new surface.
The Firm Owns Every Answer
The opinion treats AI much the way the rules treat a nonlawyer assistant. Your firm supervises it. Your firm answers for what it says.
So somebody has to read the transcripts. Pull a sample every month. Look for answers that drifted toward advice, promised a result, or invented a fact about the firm.
Set the limits in the settings rather than in a policy document. A bot told to answer only from a fixed list of firm facts will wander far less than one left to improvise.
What Good Intake Actually Looks Like
None of this argues against chat. A law firm chatbot answering at eleven at night wins matters that a voicemail loses.
The useful version is narrow. It says what it is. Then it warns about confidentiality. After that it handles hours, location, practice areas, and fees. Finally it books a consultation and gets out of the way.
- Hours, location, parking, and how to reach a human quickly.
- Whether the firm handles this kind of matter at all.
- How a consultation works and what it costs.
- A booking link, or a callback promise with a real timeframe.
Check Yours This Week
Open your own site and talk to the widget as a stranger would. Describe a problem. Ask what you should do. See whether it holds the line or starts giving advice.
Then check your own state. Florida moved early, and other jurisdictions have published guidance since. If the firm advertises across state lines, build to the strictest rule that applies to you, and stop worrying about the rest.
One more thing while you are in there. Read the widget’s own terms, and find out where the transcripts go and how long the vendor keeps them. A stranger’s case facts sitting on someone else’s server is a question worth answering before it becomes urgent.
If you want a second opinion on how your intake path reads to a stranger, start with a free website analysis, or see how we approach web design and development.
