Someone with a real case visits your website at 11pm. They have questions, they are stressed, and they will call the first firm that responds. Your contact form says “we will get back to you within 24 hours.” A competitor’s chatbot books them a consultation in 90 seconds.

An AI chatbot on a law firm website does one job: talk to every visitor the moment they arrive, sort real cases from tire-kickers, and put consultations on your calendar while you sleep. This guide covers what these chatbots actually do, what they cost, and the ethics rules you have to follow.

What a law firm chatbot actually does

Forget the image of a clumsy popup asking “How can I help?” A proper legal intake chatbot runs a structured conversation:

1. Greets and identifies itself. “Hi, I’m the intake assistant for Smith Law. I can help you get started.”
2. Asks qualification questions. Practice area, location, timeline, and the facts that matter for your intake criteria (for a PI firm: date of accident, injuries, other party insured?).
3. Screens for conflicts and representation. Captures names to run against your conflict system and asks whether the visitor already has a lawyer.
4. Books the consultation. Offers open slots and writes directly to your calendar, then sends confirmations.
5. Escalates to a human. Anything emotional, complex, or high value routes to your intake team or an attorney, with a full transcript attached.

The point is not to replace your intake staff. It is to make sure no inquiry waits until morning, and that your staff starts each day with qualified, scheduled consultations instead of a pile of cold form fills.

What it costs

Compare that to the cost of one lost case. A single signed personal injury matter can be worth tens of thousands in fees. If the chatbot captures even one extra retained client per quarter, it has paid for itself many times over.

The hybrid model most firms should use

The strongest setup is not bot or human. It is both. The AI handles first touch: instant response, fact gathering, scheduling. Your intake team handles what needs judgment: the scared caller, the complex fact pattern, the high value case that deserves a human voice. The bot also works as your safety net on nights, weekends, and holidays when no human is available.

The ethics rules you must follow

Lawyers cannot treat a chatbot like an ordinary website widget. The ethics guidance is clear on several points:

• Disclose that it is AI. Florida Bar Opinion 24-1 (January 2024) says a website chatbot must tell visitors they are talking to an AI program, not a lawyer or firm employee. An undisclosed bot that feels human is a misleading-communication risk.
• No legal advice, ever. The bot gathers facts and schedules. It must not say whether someone “has a case,” estimate claim value, or answer legal questions. Anything close gets referred to a lawyer.
• Protect confidentiality. Information a prospect shares is confidential whether or not you take the case. Ask vendors where chat inputs go, who can see them, and whether they are used to train models. Get it in writing.
• Screen for existing representation. Include questions that stop or redirect the conversation if the visitor already has a lawyer.
• You are responsible for what it says. If your chatbot gives misleading information, the lawyer is on the hook. Review transcripts regularly, especially in the first month.
• Watch advertising rules. You may say you use AI, but do not claim your AI is better than other firms’ unless the claim is objectively verifiable.

7 questions to ask any chatbot vendor

1. How does the bot identify itself as AI to visitors?
2. What stops it from giving legal advice? Show me the guardrails.
3. Where is chat data stored, who can access it, and is it used for model training?
4. How does it handle conflict checks and already-represented visitors?
5. Does it integrate with our practice management system (Clio, MyCase, PracticePanther)?
6. What happens when a conversation goes beyond its scope? How fast is the human handoff?
7. What is the full pricing, including per-conversation or per-lead fees?

Mistakes that sink chatbot projects

• Launching without the disclosure. The easiest ethics violation to commit and the easiest to avoid.
• Letting it answer legal questions. One confident wrong answer about deadlines or rights can create real liability.
• No human escalation path. A bot that traps a distressed caller in a loop does brand damage.
• Never reading transcripts. Transcripts are where you catch bad answers, find new intake questions, and prove the ROI.
• Treating it as set-and-forget. Intake criteria change, practice areas shift, and the bot’s scripts should be reviewed monthly.