6ombJoin Free
voice agentcase studyRetelllaw firmAI agents

How one AI voice agent booked 453 clients for a law firm in 8 months

·Terrell Gentry·6 min read

Most law firms lose clients at the front desk. A caller dials, hits voicemail or a hold queue, hangs up, and calls the next firm on the list. That missed call was often the most ready-to-buy lead the firm would see all week, and nobody was there to catch it.

This is the story of a voice agent we call Emily, running one firm's phone line. Over 8 months she booked 453 new clients. Here's what she actually did, the numbers behind it, and how you'd build the same thing.

What Emily actually did

Emily is a voice agent, built on Retell, that answers the firm's phone the way a trained intake coordinator would. She greets the caller, figures out what they need, qualifies the matter, and books the consultation straight onto the calendar. No hold music, no "someone will call you back."

The results over 8 months:

  • 453 new-client bookings, hello to booked, without a human touching the call
  • 571 calls handled end to end
  • 96.5% self-serve rate: nearly every caller got what they needed with zero human help
  • 176 after-hours calls caught, every single one, when the office was closed and the old system would have dropped them
  • $10K to $25K per month in revenue back to the firm

That 96.5% number is the one worth sitting with. It means the human staff got pulled in on roughly 1 call out of every 30. The rest, Emily closed on her own.

Why the missed-call gap costs so much

A person calling a law firm is not browsing. They just got in a car accident, got served papers, or got a call from a collections agency. They want to talk to someone now. If they reach voicemail, they don't leave a message and wait. They call the next firm.

Those 176 after-hours calls make the point. Old setup, they're gone. Every one of them was a person with an urgent legal problem calling outside 9-to-5, exactly when a human intake team is off the clock. Emily caught all 176. Even at a modest booking rate, that's real cases the firm would never have known existed.

The boring truth is that intake is a queue, and queues are exactly what agents win on. Not clever someday features. Answering the phone, asking the right questions, and putting a time on the calendar.

The manual way vs the agent way

Here's the same job, two ways to staff it.

Human intake deskVoice agent (Emily)
Hours coveredBusiness hours only24/7, including all 176 after-hours calls
Calls at onceOne per personUnlimited in parallel
ConsistencyVaries by mood and trainingSame script, every call
After-hours leadsVoicemail, mostly lostCaught and booked
CostSalary plus benefits per seatFlat monthly, far below one hire

The point isn't that you fire your intake team. It's that a voice agent covers the nights, weekends, and overflow spikes a human desk physically can't, and it does the repetitive qualifying so your people spend time on the calls that actually need a person.

How to build your own intake voice agent

You don't need to be a developer to stand this up. The build breaks into four steps.

1. Write the intake script first

Before you touch any tool, write down exactly what your best intake person says. The greeting, the qualifying questions, what disqualifies a caller, and the moment they offer a booking. This script is 80% of the work. A vague script makes a vague agent.

2. Build the agent in Retell

Retell is the platform we use for production voice agents, and 6omb is a Gold Retell partner. You paste your script in as the agent's instructions, pick a voice, and give it a phone number. Retell handles the hard parts: listening, understanding, and talking back in real time.

3. Connect the calendar

An agent that can't book is just a fancy answering machine. Wire the agent into your scheduling tool so it can see open slots and drop a confirmed appointment on the calendar during the call. This is what turns a conversation into a booked client.

4. Test it like a real caller

Call your own agent. Be a confused caller, an angry caller, a caller with a matter you don't handle. Listen for where it stalls or says something wrong, fix the script, and call again. With newer Claude models like Sonnet 5 and Fable 5, the agent builds and checks its own logic before you do, so you stop being the quality check on every edge case.

Common pitfalls

The teams that fail with voice agents usually trip on the same things.

  • Skipping the script. Dropping a one-line prompt in and expecting magic. Write the full intake flow, questions and all.
  • No calendar integration. The agent qualifies the caller beautifully and then can't book them. Connect scheduling from day one.
  • Never testing after-hours behavior. The 176 after-hours calls are where the money hid. Make sure your agent actually answers when the office is dark.
  • Setting it and forgetting it. Read the transcripts for the first two weeks. Real callers will say things your script didn't plan for.

FAQ

How much does a voice agent like Emily cost to run?

Far less than a full-time intake hire. Pricing depends on call volume, but Emily returned $10K to $25K per month to the firm against a flat monthly run cost. Newer Claude models made these systems cheap enough for almost any business to run.

Can a voice agent really book appointments on its own?

Yes. Emily booked 453 clients hello to booked with a 96.5% self-serve rate. The key is connecting the agent to a live calendar so it can offer and confirm real times during the call.

Do I need to know how to code?

No. Platforms like Retell let you build the agent from a written script and a phone number. The hard part is writing a sharp intake script, not writing code.

What happens on calls the agent can't handle?

You set an escape hatch. About 1 in 30 of Emily's callers got routed to a human. The agent handles the routine 96.5% and hands off the genuinely unusual ones.

If you want the exact prompts, templates, and voice-agent masterclasses behind builds like Emily, that's what the free Claude Community is for. It's where builders share the setups that book real clients, and you can start with the same intake pattern this week.

About Terrell Gentry

Founder at 6omb

Terrell is the founder of 6omb and runs Claude Community, the #1 Skool community for Voice AI agents. Over 16 months his team has built 100+ AI agent systems delivering $10M+ in business value, including voice agents like Emily, which booked 453 new clients for a law firm in 8 months. He is a Y Combinator Startup School alum (SUS20) and a Gold Retell partner.

AI AgentsClaude CodeVoice AIBusiness AutomationGrowth Marketing

Join 10k+ founders going AI-first with Claude

The Claude Masterclass, 50+ copy-paste Claude Code skills, agent-building workshops, and a community actively building the same thing you are. Free for now.

Join the free community