AI Intake for Small Law Firms: What It Collects and What Lawyers Review

A potential client calls your office while you are in court. Your assistant is on the other line with a current client. The call goes to voicemail, and the caller hangs up without leaving a message. By the time anyone checks, that person has called another firm.

The same thing happens at night. Someone fills out the contact form on your website after a car accident or a bad day with a landlord. Nobody reads it until the next morning, and nobody replies until the afternoon.

AI intake is meant to close those gaps. It answers, asks the right first questions, books the consult, and hands the details to a person at the firm. It does not give legal advice, and it does not decide who you represent. This post walks through what it can collect, how booking works, what a lawyer must always review, and what to ask about conflicts and confidentiality before you turn anything on.

How AI intake works on a call or a web form

On the phone, an AI receptionist picks up when your staff cannot. It greets the caller with your firm's name, says plainly that it is an automated assistant and not a lawyer, and asks a short set of questions you approve in advance. When the call ends, it writes a summary and sends it to the person who handles intake.

On a web form or website chat, the flow is similar. The person types instead of talks, and the AI replies in plain language, asks follow-up questions, and offers consult times.

The important part is the script behind it. You decide which practice areas you take, which matters you decline, what questions get asked, and when the AI should stop and pass the person to a human. A well-built setup follows your rules every time, including the parts staff sometimes skip when the phones are busy. If you want to see how this applies to firms specifically, our page on AI for law firms covers the main uses.

What the AI can collect from a new caller

Good intake collects enough to route the matter and run a conflict check, and not much more. Before anyone at the firm has reviewed the matter, less detail is usually safer than more.

A typical first pass gathers:

• The caller's full name, phone, email, and preferred language
• How they heard about the firm
• The general type of matter (for example, a family issue, an injury, a lease dispute, a will)
• The names of the other people or businesses involved, for the conflict check
• Any deadline the caller mentions, such as a court date or a letter they received
• The county where the issue happened
• Whether they are already working with another lawyer on this matter
• The best times for a consult

Notice what is missing: the full story. The AI can say something like, "Please share just a short summary for now. A lawyer will go over the details with you once the firm has reviewed your request." That keeps sensitive facts out of the record until a person has looked at it.

If your firm serves Spanish-speaking clients, the AI can run the same intake in Spanish, with the same questions and the same limits. Many South Florida firms need that on every line.

Conflicts: collect the names, let a person run the check

Conflict checking is one of the most important reasons to be careful with intake. The AI's job here is narrow. It collects names. It does not decide whether a conflict exists.

In practice, that looks like this. The AI asks for the names of the opposing party, any related businesses, and sometimes family members or other parties involved. It records them exactly as given. Then it sends them to the person who runs conflicts in your case management system, or it flags the intake as "pending conflict check" so nobody books a substantive meeting before that check is done.

Some firms prefer the AI not book any consult until a person clears the conflict. Others let it book a tentative slot that a staff member confirms afterward. Either can work. What matters is that you choose on purpose and write it down.

Questions to raise with your own ethics counsel or your bar's ethics guidance:

What information is reasonable to collect from a prospective client before a conflict check? How should you handle a caller who turns out to be adverse to a current client? What should your automated greeting say about the call not creating an attorney-client relationship? These are judgment calls for a lawyer, not for software, and the answers shape how the AI script is written.

Confidentiality: what to ask before you turn it on

Everything a caller says to your intake system is information your firm is responsible for. Before you use any AI tool, ask the vendor, in writing, how that information is handled. Then run the answers by your attorney or malpractice carrier.

Useful questions to ask a vendor:

Where are call recordings and transcripts stored, and who can access them? Is caller data used to train any AI model, and can that be turned off? How long is data kept, and can you set your own retention period? Can you delete a record when a person asks? Who at the vendor can see your data, and is that access logged? What happens to your data if you stop using the service?

Also decide inside the firm who sees intake summaries. A summary sitting in a shared inbox that the whole office reads is a different risk from one routed only to the intake coordinator and the reviewing lawyer.

Do not accept a vague answer like "it's secure." Ask how. And do not rely on any vendor's label to tell you that you have met your ethical duties. That is a question for your attorney.

How the AI books the consult

Once the intake questions are done, the AI can offer real open times from your calendar. It only shows slots you have marked for consults, so it will not book over court dates or client meetings.

You can set rules by matter type. A probate inquiry might go to one lawyer's calendar. An injury inquiry might go to another. A matter you do not handle can get a polite decline and, if you like, a general suggestion to contact the local bar's referral service.

After booking, the AI sends a confirmation by text or email with the date, time, office address or video link, and a short list of documents to bring. It can send a reminder before the consult and offer an easy way to reschedule. If you charge for consults, it can explain that clearly before booking, using wording you approve.

This is where an AI agent differs from a basic answering service. An agent does not just take a message. It completes a task across your tools, like checking the calendar, booking, and updating your intake record. Our page on AI agents explains how that works.

What a lawyer must always review

AI can do the first pass. It should never be the last word. Draw a clear line and make sure everyone at the firm knows where it is.

A lawyer (or staff under a lawyer's supervision, as your firm decides) should always review:

Whether to take the matter. The AI can sort and flag. Acceptance is a lawyer's decision.

Conflict results. The AI collects the names. A person runs and clears the check.

Deadlines. If a caller mentions a court date, a served complaint, or a letter with a response date, that intake should be flagged as urgent and seen by a person quickly. Never let the AI tell a caller whether a deadline matters.

Anything that sounds like a request for advice. When a caller asks "Do I have a case?" or "What should I do?", the AI should say a lawyer will discuss that at the consult, and move on.

Engagement terms and fees for the matter itself. Those belong in a conversation and a written agreement, not an automated reply.

Emergencies. If a caller describes immediate danger, the AI should tell them to call emergency services and alert staff right away.

How to tell it is working

Give it a few checks that anyone in the office can run.

Read a sample of call summaries every week. Are they accurate? Did the AI stay within the script? Did it avoid anything that sounds like advice?

Compare booked consults to inquiries. If many people start intake but few book, look at where they drop off. Maybe a question is confusing, or the available times are too limited.

Check how quickly flagged items, like conflicts and deadlines, reach a person. If urgent intakes sit untouched, the routing needs fixing, not the AI.

Call your own line after hours, in English and in Spanish. Hear what a stranger hears.

We set up these reports for firms we work with and review the results with them every month, so the numbers are not something you have to dig for.

Your next step

If missed calls and late replies to web forms are costing your firm consults, start by seeing where they happen. Our free AI audit looks at how calls, forms, and bookings move through your office today and shows where AI intake would fit, and where a person should stay in charge. You get a fixed quote afterward, with no guesswork.

We build intake into the tools your firm already uses, in English and Spanish, and train your team on how to review what comes in. Call (877) 229-4447 or email info@nurogrowth.com to talk it through.