Legal answering service: how law firms answer every call safely

What does a legal answering service do?

A legal answering service answers your firm's phone in your firm's name. It follows a script you approve. Most services offer these tasks.

• Screen new callers: It asks the type of matter, the location and how urgent it is, then decides if the caller fits your practice.
• Run intake: It collects set details, such as the caller's name, contact details and the names of other parties.
• Book consultations: It puts qualified leads into your calendar.
• Take messages for current clients: It passes messages to the right lawyer or paralegal.
• Route urgent calls: It transfers calls such as a new arrest or a court deadline to the lawyer on duty.
• Take payments: Some services take retainer or invoice payments over the phone.

How should intake protect confidentiality?

Under the ABA Model Rules, a person who talks to a firm about hiring it is a "prospective client." Model Rule 1.18 covers that person. It says a lawyer who learns information from a prospective client must not use or reveal it, even if the firm never takes the case. Your answering service is part of how the firm receives that information. So it must protect it the same way your staff would.

In practice, check these points with any vendor.

• Confidentiality terms: The contract should say the vendor keeps all call content confidential and uses it only to serve your firm.
• Who can hear calls: Ask who can listen to recordings and read transcripts, and how long the vendor keeps them.
• Training on your data: Ask in writing whether the vendor, or any AI company it uses, trains its models on your calls.
• Where data goes: Ask how intake notes reach your practice management system, and whether copies stay with the vendor.

How do conflict checks fit into intake?

A conflict check makes sure the firm does not already represent someone on the other side. The order of questions matters here. Rule 1.18 deals with "significantly harmful" information from a prospective client. A lawyer who receives it may be barred from acting against that person in the same or a related matter. That bar can spread to the whole firm. Rule 1.18 also gives an exception when the lawyer took reasonable steps to limit how much disqualifying information they received, along with screening and notice.

So a good intake script works in two stages.

1. Stage one: Collect only the caller's name, the names of the other parties and the basic type of matter. Then run the conflict check.
2. Stage two: Only after the check clears, collect the facts of the case and book the consultation.

Smith.ai, for example, sells conflict checks as an add-on for $0.50 per call on its live plans, as of October 2026. If you use AI intake, set it up to follow the same two stages. The AI should send the party names to your system or your staff for the check before it asks for the story.

What does ABA Formal Opinion 512 mean for an AI receptionist?

The ABA issued Formal Opinion 512 on July 29, 2024. It was the ABA ethics committee's first formal opinion on generative AI in legal practice. Its topics include competence, confidentiality, communication with clients and fees.

On confidentiality, the opinion says lawyers should get a client's informed consent before putting information about the representation into a self-learning AI tool. A self-learning tool is one that can keep and reuse what you type into it. The ABA litigation section notes that this duty also reaches prospective clients through Rule 1.18. It also notes that a line of boilerplate in an engagement letter is not enough.

The opinion is mainly about lawyers using AI in their own work. Still, it gives a firm clear questions to ask of an AI receptionist.

• Does the AI keep or reuse what callers say? If it does, the risk under the opinion goes up a lot.
• Do you understand how the tool works? The competence duty means a lawyer should know what the tool does with call data.
• Does the caller know they are talking to AI? Being open about it supports your duty to communicate clearly.

Your own state bar may have its own guidance too. Read it before you choose a vendor. For a wider view of AI in legal work, see my guide to AI for law firms.

How much does a legal answering service cost?

These are starting prices from each vendor's own pricing page, as of October 2026.

Service	Type	Starting price	Legal features on the pricing page
Smith.ai	Live	$300 a month for 30 calls	Intake, lead screening, conflict checks ($0.50 a call), Clio and LawPay support
Smith.ai	AI	$150 a month for 75 calls (Pro)	Native Clio integration, AI scheduling
Ruby	Live	$250 a month for 50 minutes	24/7 coverage, optional AI add-ons
Posh	Live	$65 a month plus $2.30 a minute	24/7 coverage, with a separate attorney answering page

Smith.ai's live plans also list extended intake for $1.50 per call and an option to accept collect calls. Its page notes that attorneys who work with inmates often need that. Ruby and Posh both charge by the minute, so long intake calls cost more there.

Think about cost in terms of cases. If one signed case is worth thousands of dollars to your firm, a service that books even one extra case a month can pay for itself. Track how many booked consultations the service produces in the first three months.

Should a law firm use live agents or AI?

Live agents suit practices with emotional calls, such as family law, criminal defense and personal injury. A trained person can calm a caller, hear what matters and handle a call that goes off script.

AI suits firms with high volumes of simple calls. Examples are booking calls for estate planning, traffic tickets or immigration consults. AI answers at once, at any hour, and can take many calls together. It needs strict rules for intake order and for what it must never say.

Neither option should give legal advice. The script should say plainly that the receptionist cannot advise on the caller's case, and that a lawyer will follow up.

What phone rules apply to law firm calls?

Recording rules vary by state. Federal law allows recording with one party's consent. Justia's survey lists several states that require consent from everyone on the call, including California, Florida and Pennsylvania. Tell every caller at the start that the call may be recorded.

If the AI makes outbound calls, the FCC's February 2024 ruling applies. It says AI-generated voices are "artificial" voices under the Telephone Consumer Protection Act. So keep AI voice calls for follow-ups that callers asked for, and check consent before any marketing call.

What mistakes do firms make with legal intake?

The risk in legal intake is rarely the receptionist itself, it is the order and the paper trail around it. A few mistakes come up again and again.

• Collecting the full story before the conflict check. If intake asks for case facts before party names are cleared, the firm may have already received disqualifying information under Rule 1.18, even if it never takes the case. The order matters more than almost anything else in this guide.
• Treating marketing language as a confidentiality guarantee. A vendor's page that says calls are secure or confidential is not the same as a written contract term saying the vendor, and any AI model behind it, does not train on or reuse your firm's call data. Get it in writing, specifically.
• Forgetting that disqualification can spread to the whole firm. A mistake on a single intake call is not just that lawyer's problem. Under Rule 1.18, the information a prospective client shares can bar the entire firm from a matter, which is why the two-stage order matters at the intake level, not just for the lawyer who eventually takes the call.
• Not asking who can actually hear the calls. Recordings and transcripts sitting with a vendor are still firm confidential information. Know exactly who inside the vendor, and inside any AI company behind it, can access them, and for how long.
• Assuming a generic AI tool is fine for intake. Formal Opinion 512's competence duty means a lawyer should actually understand what a tool does with call data before firm intake runs through it, not assume a popular AI product is automatically appropriate for privileged conversations.

When is a custom AI receptionist right for a law firm?

A ready-made legal service fits most small firms. Smith.ai, for example, already connects with Clio, a common practice management tool.

A custom build fits a firm in these cases:

• Your intake differs by practice area: Each area needs its own questions and its own qualifying rules.
• Your conflict check must run on its own: The AI checks party names against your own database during the call.
• You use several systems: Intake must update your case software, your CRM and your calendar at once.
• You need full control of call data: Some firms want every recording and transcript in their own accounts, with no reuse by any vendor.

My guide to law firm management software covers the systems an AI receptionist usually connects to.

How does Blur Studio approach AI intake for law firms?

I map the firm's intake by practice area first, since each one needs its own questions and its own qualifying rules. I set the two-stage order so party names come before case facts, decide exactly what the AI must refuse to discuss, and decide which calls go straight to a person.

Then I connect it to the firm's own case software and calendar, and I choose AI providers whose terms say in writing that they do not train on the firm's data. The firm owns the code and the accounts, so it controls every recording and every note, not a vendor.