AI case study

Public Employee Benefits AdministratorsCustomer support automation

QA sampled 50 conversations—most went unscored. Four agents now route and score every interaction, exposing a routing failure no one saw.

Published

Key results

Resolution Time Reduction
9%
Response Time Reduction
~6%
Interactions Evaluated
983
vs ~50 per manual evaluation cycle

Result highlights

Unlock 5 result highlights

The story

Context

A benefits administrator with roughly 20 employees serving more than 125,000 union members, including 911 dispatch operators who work around the clock, administering medical, dental, vision, pet, legal, identity protection, and supplemental insurance plans for labor unions and associations across the United States.

Challenge

Five customer service representatives handled up to 450 emails and 600 calls per week during peak periods, manually interpreting each inquiry,...

Solution
Unlock full story

Scope & timeline

  • ~1 month from UAT to production deployment

Quotes

Unlock 6 more quotes

The company

Public Employee Benefits Administrators logo

Public Employee Benefits Administrators

peba.com

Benefits administration for public employee unions and associations.

IndustryInsurance
LocationJacksonville, FL, USA
Employees11-50
Founded2013

The vendor

Enterprise software, cloud infrastructure, and consumer electronics platform.

IndustrySoftware & Platforms
LocationRedmond, WA, USA
Employees100K+
Founded1975

Use case

Public Employee Benefits Administrators's Customer support automation is part of this use case:

Agentic Customer Service
70 case studies(+33% YoY)
Proven impact?
LowModerateVery Strong
8.8Very strong
3.5Moderatewithin Insurance
6.1Strongwithin Customer Service

Similar Case Studies

Related implementations across industries and use cases

76 AI case studies in Agentic Customer Service

68 AI case studies in Insurance

475 AI case studies in Customer Service