Guide

AI voice agents and chatbots vs a trained intake specialist

AI tools now answer phones and chat windows for law firms. The decision is not human or machine. It is which job each should do, and what stays with the attorney.
Last reviewed October 29, 2026 · By Dr. Jeffrey Scott Stanton, DCH

Start with the job, not the tool

Intake is several jobs: answering, sorting, collecting facts, explaining the fee and process, asking for the consultation, following up and logging. Tools can do some of these. The ones that decide whether a caller books are the conversation, the fee and the ask.

This page is a framework. It names no vendor and makes no claim about how any product performs. Test any tool on your own calls.

A way to compare

Coverage. A tool can answer every call at any hour. A person cannot. The question is what happens after it answers.
Conversation. Does the caller feel heard? Does the tool handle upset, confused or frightened callers, interruptions and accents? Listen to full calls, not demos.
The fee and the ask. Can it state the fee as the firm sets it, and ask for a time, and respond to "I need to think about it"?
Hand-off. When does it pass to a human, how fast and with what information?
Follow-up. Who contacts the caller who did not book, and when?
Cost. Compare cost per booked consultation using your own numbers, not cost per call answered.

The legal-advice boundary applies to AI tools

A tool that answers a caller's question about their case is giving legal information to a person who may rely on it. The firm remains responsible for what its tools say. Decide in advance what the tool may never say, and test it with the questions callers actually ask: do I have a case, what should I do, what is it worth.

See the line between intake and the attorney. The same rules apply to a machine as to a person.

Questions to ask any vendor

What does the tool say when a caller asks whether they have a case?
Where are calls and transcripts stored, who can access them and for how long?
How does the tool handle consent and notice for recording and for being an automated system?
Which parts of the conversation are scripted and which are generated?
How are mistakes found and corrected, and who reviews calls?
What happens when the tool cannot understand the caller?
How does the firm export its data if it leaves?
What do other firms in the same practice area say, and can the firm speak to them directly?

Where each fits

Many firms will use a mix: software for coverage and sorting, trained people for the conversations that book. Score both with the same scorecard. If the tool scores worse than a trained person on the fee and the ask, it is a screening tool, not an intake desk.

Keep reading

Scope, plainly: everything on this site concerns the conversation that turns an inquiry into an engagement. Nothing here is legal advice, nothing touches the legal work or the attorney's judgment, and any wording a firm adopts is reviewed by the firm against its jurisdiction's rules on attorney advertising, solicitation and fee communication before it is used.

About the Author

Dr. Jeffrey Scott Stanton, DCH is a behavioral strategist and the founder and Chief Learning Officer of R2R Diagnostic Group. He is the former EVP of Learning & Development at Douglas Elliman Real Estate, where he trained sales professionals on the conversations that turn an inquiry into a commitment. He applies the same method, the R2R Assessment and a 14-item call-review scorecard, to law firm intake teams. The work is the conversation only; legal questions and judgment stay with the attorney.

The method, the scripts and the scorecard are on how it works, the script template and the scorecard guide.

Start With the Calls

Want this scored on your own calls?

Send a handful of recordings or transcripts. You get a scored review by person, and the one moment to fix first.

Start With the Audit or book a conversation