How a 14-Person Insurance Agency Answered 877 Calls in 30 Days Without Hiring a Single Receptionist

The Problem: Too Many Calls, Not Enough People Who Want the Job

Libardi Insurance runs lean - about 14 people covering a book of business that's roughly 80% commercial. Co-owner Ben Libardi described the daily reality bluntly during an early sales conversation: dozens of inbound calls a day, a coverage gap in the mornings before the office opened, no safety net after hours, and a growing number of Spanish-speaking customers with nobody dedicated to serve them in-language.

Hiring wasn't a simple fix. Ben and his team were fielding job applicants asking for $70,000–$140,000 a year for entry-level, unlicensed customer service work - candidates who still needed training on day one. On top of that, spam and solicitation calls were eating into staff time and patience.

As Ben put it, half-joking but not really:

Let him pay the $65,000 for the receptionist, with full benefits.

Ben

Co-Owner

The agency didn't need more headcount. It needed its front door to actually work.

‍

Why Libardi Chose an Insurance-Specific AI

Libardi evaluated its options carefully, including outsourced virtual assistants and other automation vendors, before settling on Sonant. The deciding factor wasn't a feature checklist - it was fit. As Ben explained afterward:

The fact that it's insurance-specific, insurance agent-specific, is a huge plus, because she gets the language without making it complicated for the customer. I can't imagine the AI really being able to successfully advocate those calls the way she does.

Ben

Co-Owner

‍

Because Sonant is built exclusively for P&C agencies, the AI receptionist - which Libardi's team nicknamed "Libby" - already understood the difference between a personal auto quote and a commercial certificate-of-insurance request, without months of custom training. It could route calls based on who was asking for what, transfer warm to the right person when someone was available, and hand off cleanly to voicemail or a callback queue when they weren't.

Anthony Libardi, who runs the agency's South Carolina office, had a similar reaction during the initial demo:

It seems very realistic, probably more realistic than I anticipated. It's not like a generic AI bot, but it looks pretty good on my end.

Anthony Libardi

‍

From a 3-Hour Test to a Permanent Fixture

Libardi's rollout tells its own story. The plan was cautious: turn the AI receptionist on for a few hours on day one, see how it goes, then decide. As Ben later admitted on a follow-up call:

We told you we only intended on putting it on that first day for like three hours. We ended up never turning it off. That goes to show you we're confident in it.

Ben

Co-Owner

‍

That confidence shows up in the usage data. Over a recent 30-day stretch, the system handled 877 total calls - every single one answered, with zero routed to a dead end. Calls that came in outside business hours or during the agency's known coverage gap (9:00–9:45am) were collected, summarized, and routed automatically instead of disappearing into an empty inbox.