Home/The Retainer/AI and intake
The Retainer · AI and intake

When Callers Need a Person: Where AI Intake Stops

AI intake can answer every call and book the consult. It shouldn't give advice, handle safety calls alone or argue with someone who wants a person.

MAY 6, 20255 MIN READBY THE INTAKEGENIUS TEAM
Three pairs of hands held together
Photo: Unsplash

If you're thinking about AI intake, you've probably had the same worry most attorneys have. What happens when the caller needs a human? What if the AI says the wrong thing to someone in a crisis, or a caller just wants a person and gets stuck talking to software?

Those are fair questions. AI intake is good at a lot of things, and it has real limits. Here's an honest look at where it stops, and how to set it up so those calls reach a person.

What AI intake does well

Start with what it's for. A well-built AI intake agent can:

  • Answer every call, at any hour, without a hold queue.
  • Ask the same practice-area questions every time, in the same order.
  • Capture the details your attorney needs in a clean, readable record.
  • Book the consultation on the call and confirm it by text.

That covers most intake calls at most small firms. A person calling about a will, a divorce, a bankruptcy or a green card mostly wants to know you handle it and get on the calendar.

But "most" isn't all. The rest is where you need a plan.

Where AI intake should stop

This one's simple. AI intake doesn't give legal advice. Neither should your human intake staff.

Callers ask anyway. "Should I sign this?" "Will I lose custody?" "Can they take my house?" The right answer from intake, AI or human, is the same: that's a question for the attorney, and here's when you can talk to them.

"That's a really good question for the attorney. I can't advise you on it, but I can get you on their calendar so you can ask them directly."

If an AI agent is ever pushed past that line, the setup is wrong. Make sure the script holds firm.

Safety and emergencies

Some callers aren't safe. A family law caller who says their partner is in the house. Someone who mentions hurting themselves. A probate caller who's clearly in a mental health crisis.

AI intake shouldn't try to handle these alone. Its job is to recognize the signals your firm defines, tell the caller to contact emergency services when there's immediate danger, and route the call to a person at your firm right away. Your firm decides who that person is for each kind of call. We go deeper on this in can AI handle distressed callers and in our guide to safe intake for domestic violence callers.

Complex judgment calls

Some decisions need a human brain, and usually an attorney's. Examples:

  • Whether to take a case that's outside your normal practice.
  • A possible conflict that isn't clear-cut.
  • A fee negotiation or an unusual payment arrangement.
  • A caller with three overlapping legal problems.
  • Anything where the "right" answer depends on judgment, not a script.

AI intake can collect the details and book the consult. It shouldn't decide these. Write down which situations always go to a person for review before anyone commits to anything.

Upset callers who ask for a person

Some callers don't want to talk to AI. Maybe they're angry, maybe they're grieving, maybe they just prefer people. If a caller asks for a human, they should get one.

Don't let an AI agent argue, deflect or keep pushing the intake questions. That turns a lead into a bad review.

"Of course. Let me get you to someone on our team."

Then make sure your rules name someone who'll actually pick up. A handoff to nobody is worse than voicemail.

Should callers know it's AI

We think yes. People can usually tell, and finding out halfway through feels like a trick. A short, plain line at the start of the call is enough.

The rules around AI disclosure are still changing and can vary by state, so check what applies to your firm. But even where it's not required, being upfront tends to build trust instead of costing it.

The handoff is the whole game

AI intake works when the handoff to people is clean. It fails when the AI collects a lead and nobody owns what happens next.

For every situation where AI stops, define:

  • The trigger. What the caller says or does that means "a person needs this."
  • What the AI says. Exact wording, approved by an attorney.
  • Who gets it. A named person or role, not "the team."
  • When. Right away for safety issues and callers who ask for a person, and who covers which hours.
  • Where it's logged. One place, so nothing gets lost.

If you can't fill in all five for a situation, you're not ready to let AI handle that call type yet.

When a person should answer first

For some firms, a few call types should go to a person from the start. A firm with a heavy domestic violence caseload, for example, might want a trained human on more of those calls. That's a reasonable choice.

That's where a hybrid setup or a dedicated intake person helps. IntakeGenius offers an optional add-on: a full-time intake specialist trained by our team, who works alongside the AI on the calls that need a human. We cover how the two fit together in hybrid AI and human intake for law firms.

Where IntakeGenius fits

Our AI voice agent answers every call, runs your intake and books the consult, with the escalation wording and triggers your attorney approves. When a call needs a person, it routes the call to someone at your firm under those rules, and the lead lands in IntakeCRM with the details. If you want a human on those calls, our intake specialist add-on is there.

Want to test it with your hardest call types? Book a 15-minute demo. Bring your worst-case scenarios, we'll skip the slides.

Intake. Solved.

Let intake run itself.

See how IntakeGenius handles intake end to end.